Simplicial-based techniques for multi-resolution volume visualization: an overview
Rita Borgo1 , Paolo Cignoni1 , and Roberto Scopigno1
Istituto di Scienza e Tecnologia dell’Informazione (ISTI), Consiglio Nazionale delle Ricerche, via G. Moruzzi 1, 5614 Pisa, Italy. Email: borgo, cignoni@https://www.doczj.com/doc/5a1267658.html,r.it, roberto.scopigno@https://www.doczj.com/doc/5a1267658.html,r.it
1 Introduction
A number of approaches exist to support multi-resolution management: na¨ ?ve sub sampling, wavelet techniques, methods based on hierarchical space subdivisions (e.g. octrees), methods based on simplicial decomposition. The term multi-resolution is often used to indicate either discrete or continuous level of detail (LOD) representation. We will cover mostly the second aspect, and therefore we point our attention to those methods that allow to manage selective re?nements (or the inverse operation, i.e. selective coarsening) in a dynamic manner, according to the interactive requests of the application. Simplicial meshes have been often used in the visualization of volume dataset. The simplicity of the basic cell allows to easily manage isosurface extraction (?eld is linearly interpolated, no ambiguity) and to implement in an e?cient manner direct volume rendering (DVR) solutions. Moreover, tetrahedral-based DVR solution can now be implemented using o?-the-shelf graphics hardware, gaining impressive speed-ups with respect to software solutions. Therefore, simplicial decompositions have been often considered in the design of multi-resolution methods, not only because easy to render but also because they easily adapt to di?erent shapes or to the data ?eld structure/topology. This paper presents an overview and a comparison of the different approaches based on multi-resolution simplicial decomposition, which have been proposed in the context of volume visualization. We subdivide the existing methods in two main classes, which depend on the re?nement kernel used to manage the selective re?nement/coarsening: regular or irregular. Regular techniques starts from a coarse irregular base domain and apply recursive regular re?nement, resulting in large meshes organized as uniform grid patches. Irregular techniques are independent from the topology of the underlying mesh and as consequence during the re?nement procedure new vertices can be inserted at more convenient locations instead of prede?ned positions; not being forced to follow a regular subdivision scheme irregular techniques result to be more ?exible and suitable to resolve complex geometric features and geometry changes.
2
Rita Borgo, Paolo Cignoni, and Roberto Scopigno
2 Using Irregular Re?nement Techniques
As introduced in the previous section, multiresolution models for generic tetrahedral meshes has evolved to manage set of level of details in a more ?exible, e?cient and compact way. The main idea behind these methods is to exploit, in some way, the information that can be collected during the simpli?cation of a volume dataset. The assumption is that we use an iterative simpli?cation algorithm that progressively reduces/re?nes the dataset by means of small local operations. This sequence of small modi?cations is organized in a structure that encodes all the temporal dependencies among themselves and makes it possible to apply backward the small modi?cations in a di?erent order. In this way is possible, for example, selectively re?ne, or simplify, only certain portions of the domain, according to the user needs. These techniques are therefore strongly related with multiresolution and simpli?cation techniques for generic three-dimensional surface meshes where those ideas were ?rstly developed (see, e.g., [10] for a survey of surface multiresolution). For this reason in the next section we shortly review the existing techniques for the simpli?cation of a tetrahedral complex. 2.1 Simpli?cation of a simplicial complex One of the ?rst approaches to the construction of a simpli?ed representation of a tetrahedral dataset was proposed in [3]; it exploits a basic coarse-to?ne re?nement strategy, an early technique developed in the two-dimensional case and widely used for approximating natural terrains [7, 11]. An on-line algorithm for Delaunay tetrahedralization is used together with a selection criterion to re?ne an existing Delaunay mesh by inserting one vertex at a time. The selection strategy at each iteration is aimed to re?ne the tetrahedron that causes the maximum warping/error in the current approximation: this is obtained by selecting the datum vmax corresponding to the maximum error as a new vertex. After a point is added to the dataset, the tetrahedron that contains it is split and a sequence of ?ipping actions (see Fig. 1) is performed until the resulting mesh satis?es the Delaunay criterion. This re?nement procedure always converges since the number of points in V is ?nite; total accuracy is warranted when all of them are inserted as vertices of Σ. This approach is limited to datasets whose domain is convex, because the result of a Delaunay tetrahedralization is always convex. We have proposed the extension of this approach in [5] to deal also with non-convex curvilinear datasets; in this case the Delaunay tetrahedralization is computed in the computational domain (i.e. the underlying grid), while its image through lifting gives the corresponding mesh in the physical domain. The vertex insertion method described above is di?cult to adapt to the case of generic non-convex irregular datasets. Major di?culties arise in ?nding
Title Suppressed Due to Excessive Length
3
Fig. 1. A tetrahedra split, due to a new vertex selection and insertion in the mesh (top image); the two classes of tetrahedra ?ip actions: the 2 to 3 ?ip, which produces three cells out of two (center image); the 3 to 2 ?ip, needed when the two tetrahedra present a non convex union (a), and a third cell (b) has to be included in the ?ip action.
an initial coarse mesh to approximate the original domain ? of the dataset and in the estimation of warping. Delaunay triangulation is not applicable to non-convex polyhedra; moreover even if we have an approximation of the boundary of the starting domain ?nding a tetrahedralization of this polyhedron, without adding new points, is an NP-complete problem [19]. Experience in the approximation of non-convex surfaces through 2D triangular meshes suggests that a decimation technique might be more appropriate to the case of non-convex irregular volume datasets (see, for example, [1,12,20]). Similarly for the 2D surface case in the decimation approach we start from the reference mesh Γ , and vertices are iteratively discarded as long as the error introduced by removing them does not exceed a given a accuracy threshold. Gross and Staadt [21] present a simpli?cation technique based on collapsing an edge to an arbitrary interior point, and propose various cost functions to drive the collapsing process. Cignoni et al. [5] propose an algorithm based on collapsing an edge to one of its extreme vertices (called half-edge collapse see Fig. 2), in which the simpli?cation process is driven by a combination of the geometric error introduced in simplifying the shape of the domain and of the error introduced in approximating the scalar ?eld with fewer points. This approach has been extended in [2] by de?ning a framework for the uni?ed management of the two errors (related to geometric domain and scalar ?eld) and by proposing some techniques for an e?cient evaluation and forecast of such errors.
4
Rita Borgo, Paolo Cignoni, and Roberto Scopigno
a
b
Fig. 2. Half-Edge collapse in 2D: (a) a valid collapse; (b) an inconsistent collapse.
In fact, selecting a vertex to be removed involves an estimation of the amount of ?eld error and geometric domain error due to the removal: the criterion usually adopted is that the vertex causing the smallest increase in error/warping should be selected at each iteration. An exact estimation of the change in error and warping can be obtained by simulating deletion of all vertices in the current mesh. This would be computationally expensive, since each vertex has 24 incident tetrahedra on average. This may involve relocating many points lying inside such tetrahedra. We prefer to use heuristics to estimate apriori how a vertex removal a?ects error and warping. Such an estimation is computed for all vertices before decimation starts, and it is updated for a vertex each time one or more of its incident tetrahedra change. Trotts et al. [22] perform half-edge collapse as well. They control the quality of the simpli?ed mesh by estimating the deviation of the simpli?ed scalar ?eld from the original one, and by predicting increase in deviation caused by a collapse. They also provide a mechanism to bound the deformation of the domain boundary. Another approach for the simpli?cation of generic simplicial complexes, called Progressive Simplicial Complex (PSC), has been proposed by Popovic and Hoppe [16], as an extension of the Progressive Meshes (PM) model [12]. The PM models are based on the simpli?cation of a mesh with a sequence of edgecollapse transformations, this sequence of operations is encoded in the PM structure. Given a PM, a mesh can be reconstructed by applying, in the right order, a series of vertex-split transformations (the reverse of edge-collapse). The PSC codi?es in a similar manner the sequence of more general edgecollapse transformations. It should be noted that while the PSC are quite general, they have been conceived for the management of possibly degenerated 2D surfaces rather than simplicial complexes representing a volume dataset, so the conditions of legality of a sequence of vertex-split operation of a generic complex are not speci?ed, and the problem of evaluating the approximation error introduced in the volume ?eld representation is not considered. 2.2 From simpli?cation to Multiresolution Models In this section we introduce the framework of multiresolution simplicial models (MSM) introduced by De Floriani, Puppo and Magillo [8] as a multidi-
Title Suppressed Due to Excessive Length
5
mensional extension of the two dimensional structure described by Puppo in [17]. De?nitions Let S be a ?nite set. A partial order on S is an antisymmetric and transitive relation < on its elements. A pair (S, <) is called a partially ordered set (poset). For every s, s ∈ S with s < s we mean that s < s and s such that s < s < s . A subset S ? S is called a lower set if ?s ∈ S , s < s ? s ∈ S . Intuitively S is a lower set if it contains all the elements that precede each of its elements. The algebraic structure of a poset (S, <) can be described by a direct acyclic graph (DAG), where nodes represent elements of S and arcs encode the < relation. For any s ∈ S the set Ss = {s ∈ S|s ≤ s} is the smallest lower set containing s and it is called ? the down-closure of s. The sub-closure of s is de?ned as Ss = Ss \{s}. n We call any ?nite set of d-simplexes in IE a d-set; a regular d-simplicial complex Σ is completely characterized by the collection of its d-simplexes i.e. by the d-set associated with Σ. When managing a collection of representation of the same complex at different resolutions, as done in the previous section, we need a measure of the error we commit. With μ(σ) we denote a function μ : Σ → [0, 1] measuring this error, μ(0) means exact representation. With μ(Σ) we denote the maximum error among all the tetrahedra of Σ: maxσ∈Σ μ(σ). Operators. We
Σ8
Σ6
Σ7
Σ4
Σ5
Σ1
Σ2
Σ3
Σ0
Fig. 3. The DAG describing a simple two-dimensional MSM.
de?ne two operators on d-sets: the interference operator ? and the combination operator ⊕. Both operators take two d-sets as arguments and produce a
6
Rita Borgo, Paolo Cignoni, and Roberto Scopigno
d-set. The interference operator of two d-sets is de?ned as: Σi ? Σj = {σ ∈ Σi |?σ ∈ Σj , i(σ) ∩ σ = ?} In other words, the interference of two d-sets Σi , Σj is the set of the simplexes of Σi that have a proper intersection with some simplexes of Σj . The combination operator of two d-set is de?ned as: Σi ⊕ Σj = Σi \(Σi ? Σj ) ∪ Σj In other words, the combination of two d sets Σi , Σj is the set of the simplexes that can be obtained by adding to Σj the simplexes of Σi not interfering with Σj . If Σi ⊕ Σj is a d-simplicial complex and ?(Σi ⊕ Σj ) = ?(Σi ) ∪ ?(Σj ), then the complex Σj it is said to be compatible over Σi . Let [Σ0 , . . . , Σk ] be a sequence of d-complexes. the combination ⊕k Σk is i=0 de?ned as: – if k = 0 then ⊕0 Σk = Σ0 i=0 – if k > 1 then ⊕k Σk = (⊕k?1 Σk ) ⊕ Σk i=0 i=0 Multiresolution Simplicial Model A Multiresolution Simplicial Model (MSM) on ? is a poset (S, <), where S = {Σ0 , . . . , Σh } is a set of d-complexes and < is a partial order on S satisfying the following conditions: 1. ?(Σ0 ) = ?, 2. ?i, j = 0..h, i = j,, a) Σi < Σj ? Σi ? Σj = ? b) Σi ? Σj = ? ? Σi is in relation with Σj 3. the sequence [Σ0 , . . . , Σh ] of all complexes of S de?nes a consistent order w.r.t. relation < and [Σ0 , . . . , Σh ] is a compatible sequence. The meaning of the second condition becomes clearer if we consider the DAG encoding relation < for the set S: – if two complexes Σi and Σj are connected by an arc, they are interfering; – if two complexes Σi and Σj interfere then they are connected through a path. The elements of S are called components or fragments; intuitively the fragments describe a portion of the domain ? at a certain resolution. For example, if we think to an iterative re?nement procedure on a simplicial mesh, the set of simplexes, derived from the substitution of a complex with a more re?ned one, can be considered as a fragment that is combined over the existing complex. Combining a lower set of S give us a complete description of ? at various resolutions. The following properties holds for MSM’s: Lemma 1. Σ0 is unique minimum element of (S, <).
Title Suppressed Due to Excessive Length
7
In other words Σ0 is the starting coarsest complex. Given any subset S ? S the total order of its elements consistent with the sequence [Σ0 , . . . , Σh ] is called the default order of S . Lemma 2. The default order of any lower set S ? S is a consistent order. Lemma 3. In a MSM (S, <), the combination of a lower set S ? S is independent of the speci?c consistent order. Since the combination obtained from any consistent order is unique, it will simply called the combination of S and denoted with ⊕S. Let Σi be a com? ponent of (S, <); the combination of the subclosure SΣ is called the support of Σi ; the set of the simplices of the support that are interfering with Σi , ? (⊕SΣ ) ? Σi is called the ?oor of Σi . The following de?nitions permit us to consider a particular class of MSM’s where the order relations provide control over the size in terms of number of simplexes. A MSM (S, <) is increasing if and only if for every pair of lower sets S , S holds: (S ? S ? |⊕S | < |⊕S |. Similarly is de?ned the concept of decreasing MSM; an increasing or decreasing MSM is said monotone. In other words a MSM is increasing (decreasing) if and only if the size of each fragment is larger (smaller) than the size of its ?oor. In Figure 3 is shown a simple multiresolution simplicial model for the two dimensional case; the arrows in the ?gure correspond to the relation < . The fragments of a MSM can be used to build di?erent triangulations of the domain ? through the paste operator. The intuitive meaning of the < relation is of dependence between the pasting of the fragments: if we use a fragment Σi in a triangulation then all the fragments Σj < Σi must also be used. In Figure 4 is shown the triangulation resulting from the pasting/combination of a subset S of fragments in a consistent order Σ0 , Σ2 , Σ3 , Σ4 , Σ7 ; note that any other consistent order of pasting of S (like Σ0 , Σ3 , Σ2 , Σ4 , Σ7 ) builds the same triangulations. 2.3 MSM for Volume Visualization Each simpli?cation algorithm described in Section 2.1 can be used to build a MSM. Both a decimation or a re?nement algorithm for simplifying a tetrahedral complex can be regarded as producing an “historical” sequence of tetrahedra, namely all tetrahedra that appear in the current mesh Σ during its construction. An historical sequence can be also viewed as the sequence of all subdivisions of the whole domain that are obtained through changes, or as an initial subdivision plus a sequence of fragments re?ecting the local changes iteratively done to the mesh, i.e. subdivisions of portions of the domain, which can be partially overlapping and are pasted one above the other to update the existing structure.
8
Rita Borgo, Paolo Cignoni, and Roberto Scopigno
Σ8
Σ6
Σ7
Σ4
Σ5
Σ0⊕Σ2⊕Σ3⊕Σ4⊕Σ7
Σ1
Σ2
Σ3
Σ0
Fig. 4. A subset S ? S combined in a consistent order builds a triangulation
For example, if we follow the re?nement heuristics, the minimum of the poset is the starting coarse triangulation; when we insert a new point vi in the complex the new tetrahedra that are built form a new fragment Σi ; the ?oor of this fragment is constituted by the tetrahedra that were destroyed by the insertion of vi . Following the MSM all these fragments, represented by a tetrahedral complex covering a small part of the whole domain ?, are arranged in a poset where the order relation between fragments is dependent on their interferences in 3D space. The minimum fragment Σ0 , the coarsest representation of our mesh, has an empty ?oor. Similarly all the triangles on the top of S, representing the dataset at its full resolution, have no upper fragments. A simple data structure to encode a generic MSM was presented in [6]; in [4] a much more compact data structure has been proposed for a three-dimensional tetrahedral MSM built based on a sequence of general edge collapses; this structure, customized to the needs of volume visualization, requires three times less storage space with respect to a simple indexed data structure encoding the original mesh at full resolution, and 5.5 times less space than a data structure for the original mesh encoding both connectivity and adjacency information (as required, e.g., by direct volume rendering algorithms based on cell projection).
Title Suppressed Due to Excessive Length
9
2.4 Extracting a variable resolution model The algorithm for the extraction of a variable resolution model here presented are a straightforward extension to tetrahedral complexes of the one presented in [17]. We suppose that our MSM is monotone, to extract a variable resolution model we need a boolean acceptance function c(σ) in order to decide, for a given tetrahedron, if its accuracy is su?cient or if we need to further re?ne that part of the domain. The algorithm for the extraction of a variable resolution model tries to incrementally build the desired solution by adding new fragments to the current solution. The algorithm is based on a breadth-?rst traversal of the DAG representing the MSM. The traversal starts from the coarsest fragment Σ0 , root of the DAG, and fragments above the current solution are progressively traversed and marked. The current solution is maintained as a list of tetrahedra ΣOut . For each fragment Σ that we encounter in the traversal of the tree, the following two loops are executed: – we search for fragments before Σ, still not visited and, if found, they are added to the traversal queue Q. All the fragments before Σ can be found by checking, for each tetrahedron σ ∈ Σ, if the corresponding lower fragment Lower(σ), has been marked. – for each tetrahedron σ ∈ Σ, if it satis?es the acceptance function c(σ) then σ is added to the current solution, else we add the upper fragment of σ to the traversal queue Q and mark it to be removed from the solution. The correctness of this algorithm has been proved in [6].
3 Using Regular Re?nement Techniques
Within regular subdivision the world is divided if not into equal sized voxel at least into regular shaped entity following a recurrent pattern. Simple mathematical rules withhold the basis of a subdivision scheme that is applied in a recursive manner.The adoption of a prede?ned rule to perform the subdivision introduces some constrains on the topology of the local region to be re?ned/simpli?ed, and in e?ect this type of approach adapts just to regular dataset. Initially, this approach has been proposed for terrain data represented by gridded [9], and then been extended to represent voxels dataset with a simplicial decomposition constructed via a regular subdivision rule [13, 14, 23]. Essentially the re?nement heuristic consists of a uniform recursive subdivision of the volume data, driven by a set of simple mathematical rules that guarantee a progressive update of the accuracy of the intermediate representation in an error-controlled manner. The subdivision process in general starts by subdividing the bounding box of the volume (i.e. a single hexahedral cell) in simplicial cells. The subdivision proceeds recursively and can be described either as a vertex-adding process or,
10
Rita Borgo, Paolo Cignoni, and Roberto Scopigno
analogously, as a cell subdivision process. Each step picks up a vertex from the original volume and divides the cell containing it (or a group of adjacent cells) in two (or more) simplicial cells. Because of the regular and hence predictable parametric structure of the regular re?nement, these techniques always generate meshes with bijective mapping between the coarsest levels and the ?nest levels. Therefore, going from ?ne/coarse to coarse/?ne levels correspond in executing a regular fusion/subdivision of simplices allowing for smooth and continuous changes between di?erent levels of details. Recursive subdivision schemes automatically achieve hierarchical multi-resolution representations, hence, regular re?nement techniques guaranteeing an almost everywhere regular structure allows for e?cient tree based data structures organization with higher performances on modern processors. The beauty of regular re?nement techniques basically lies in their elegant mathematical formulation and in the simplicity of the rules for generating di?erent representations. In the next paragraph we introduce in detail some of the main contribution in the ?eld of regular re?nement techniques.
3.1 Re?nement of Tetrahedral Grids Regular subdivision scheme allows for the organization of the space of the data in a hierarchical manner. As the subdivision proceeds a more detailed version of the surface is produced causing each level of the hierarchy to hold implicitly a multi-resolution organization of the dataset itself. A hierarchical organization of the dataset allows as primary consequence for a selective traversal of the representation and indeed fast construction of adaptive levels of detail. For what concern volume visualization not many attempt has been tried to accomplish good balance between e?cient subdivision techniques and easiness of implementation. Techniques for tetrahedra bisection mostly recall similar patterns of subdivision, for this reason we have individuated three principal contribution in the area of regular volume re?nement techniques based on simplicial complexes. As ?rst work we introduce the one from Plaza and Carey [15] in which re?nement and conformity are protracted together. De?nitions Skeleton. Let ? be a bounded set in Rn with non empty interior and τ an n-simplicial mesh of ?. The set skt(τ ) = {?(ti ) : ti ∈ τ } will be called skeleton or (n-1)-skeleton of τ with ? being the boundary in Rn . The skeleton of a tetrahedralization in 3D dimensions is comprised of the faces of the tetrahedra, in 2D is the set of edges of the triangles and in 1D it is the set of the points which de?ne the partition into segments. Edge bisection. Let S =< X0 , X1 , . . . , Xn > be an n-simplex in Rn , with edge < Xk , Xk > having midpoint A = (Xk + Xk )/2. Then two new simplices S1 =< X0 , . . . , Xk?1 , A, Xk+1 , . . . , Xk , . . . , Xn > and S2 =<
Title Suppressed Due to Excessive Length
3 3
11
3
1 (b) First step
2
1 2 (d) Refinement to the right
1
(a) Starting TRiangle
2
3
3
1 (c) Second step
2
1 2 (e) Refinement to the left
Fig. 5. 2D version of the 4T Rivara’s algorithm
X0 , . . . , Xk , . . . , Xk ?1 , A, Xk +1 , . . . , Xn > may be formed such that the interiors are disjoint and S = S1 ∪ S2 . This de?nes a subdivision of S by edge bisection. Re?nement of Tetrahedral Grids: Plaza and Carey Technique Plaza and Carey’s starting point is a modi?ed version of the 4T algorithm from M.C. Rivara [18]. In 2D such techniques can be seen as triangle re?nement based on a longest edge bisection criteria and works as follows. Considering a generic triangle t, the 4T algorithm divides t in 4 new triangles as in Fig. 5. The subdivision consists of two steps showed respectively in Fig. 5b-c. In the ?rst step a selection of the longest edge of the triangle t is performed and the triangle is bisected in two halves by the line joining the bisection point and the “opposite vertex”. The second step proceeds with the subdivision of the triangles produced in step 1 following the same bisectioncriteria as for the original triangle. The 4T algorithm is used by Plaza as re?ning technique of what is called “skeleton” of the mesh to be re?ned. A point worth noting is that to guarantee the conformity of the mesh Plaza induces the subdivision of all simplices that contain the bisected edge. The 3D version of the 4T algorithm applies to tetrahedral meshes in the same way the 4T algorithm applies to triangular meshes. Let T n = {τ1 < τ2 < . . . < τn } be a sequence of nested three dimensional grids with τ1 be the initial mesh and τn the ?nest mesh in the sequence, τn+1 , with τn < τn+1 , is generated applying a 2-dimensional re?nement algorithm to the skeleton of τn . The algorithm works as follows: ? Step 1: all the t ∈ τn that need to be re?ned are selected; ? Step 2: a node is added at the midpoint of each edge of each selected t;
12
Rita Borgo, Paolo Cignoni, and Roberto Scopigno
? Step 3: the conformity of each t ∈ τn is checked. If a t is non-conforming a node is added at the midpoint of the longest-edge of each non-conforming face of t and also to the midpoint of the longest edge of t; ? Step 4: each t ∈ sktτn is subdivided by the 4T algorithm of Rivara; ? Step 5: each t ∈ τn is properly subdivided,
(a) Starting mesh.
(b) Steps 1 and 2
(c) Step 3: checking conformity
(d) Step 4: skeleton disvision
(e) Step 5: division in terahedra
Fig. 6. Application of the 3D algorithm
the result produced can be seen in 6e. In Plaza’s algorithm the most di?cult case is represented by the presence of regular faces or elements, in this case further heuristics need to be formalized: the ?rst edge in which has been introduced a node for division is chosen as the longest edge of the face/element. This kind of choice avoids breaks in the re?nement area and limits the growth of the subdivision because of required conformity. The ?ve steps of the algorithm are showed in Fig. 6.
Title Suppressed Due to Excessive Length
13
As second contribution we decided to introduce a work from Zhou et al. [23]. Quite di?erent from Plaza’s algorithm, Zhou work still adopts a longest edge bisection criteria approach to perform the subdivision of the volume. Re?nement of Tetrahedral Grids: Zhou, Chen and Kaufman Technique Starting point of Zhou’s algorithm is the bounding box of the volume seen as a cube with faces parallel to a coordinate plane. The bounding box is subdivided into tetrahedra. The center of the box is added to the volume and connected to all the 8 vertices forming 6 pyramids. Each pyramid can be easily divided into two tetrahedra by connecting the diagonal of the base face generating 12 tetrahedra. After this initial step the algorithm proceed as a regular scheme for tetrahedra subdivision. Each produced tetrahedra is recursively subdivided adding a midpoint to its longest edge and connecting the new vertex with the opposite one lying on the “base” face. The tetrahedra generated by the subdivision can be grouped in three main classes: ? Class 1 cell: there is only one face parallel to a coordinate plane and there exists only one edge l of the face not parallel to any coordinate axis; ? Class 2 cell: there is only one face parallel to a coordinate plane and there exists only one edge l of the face parallel to a coordinate axis; ? Class 3 cell: there are only two faces parallel to coordinate planes (the edge that does not belong to any of these two faces is denoted by l).
A C C B A B A C D D B D
(a) Case 1
(b) Case 2
(c) Case 3
Fig. 7. Types of tetrahedra generated by the subdivision.
Figure 7 depicts the three classes just described. Edge AB of tetrahedron ABCD correspond to edge l. In each di?erent case the midpoint of corresponding edge l is selected as the dividing point. In Fig. 8 is shown the entire subdivision process. After a few steps in which the cube subdivision is performed the cycle ends with 12 tetrahedra all belonging to Class 1 (Fig. 8c), that subdivided during step 3 produce tetrahedra belonging to
14
Rita Borgo, Paolo Cignoni, and Roberto Scopigno
A C B
D
(a) Starting step A C B C
(b) Step 1 A B C
(c) Step 2
A B
D
D
D
(d) Step 3
(e) Step 4
(f) Step 5
Fig. 8. Cube subdivision by Zhou et al.
Class 2 (Fig. 8d) that subdivided again during step 4 of the subdivision process produce tetrahedra belonging to Class 3 (Fig. 8e). After the subdivision of Class 3 tetrahedra the con?guration recursively returns to Step 2. It is worth noting that each type of tetrahedra belongs to a di?erent step of the subdivision process and that the overall process has a cyclic behavior. Third and last contribution is the one from Pascucci and Borgo that introduces a subdivision schema recalling the one from Zhou et al. but that di?ers deeply in the subdivision structure. Re?nement of Tetrahedral Grids: Pascucci Technique As for Zhou et al. the starting point of Pascucci [13] techniques is the volume bounding box, a cubic cell, divided in function of its center. The center of the box is added to the volume and connected to all the 8 vertices forming 6 pyramids. Pascucci do not divides the cube into tetrahedra instead he considers the new entity formed by the six pyramids, generated by the subdivision, with the pyramids generated by adjacent subdivided cubes (Fig. 10b). This new entity corresponds to an hexahedral shaped cell characterized by a center that corresponds to the center of one of the faces of the cube. The center of the hexahedra is added to the volume and connected to all the 6 vertices dividing the cell into six tetrahedra. This six tetrahedra correspond each to the eighth part of di?erent octahedral shaped cells whose centers correspond
Title Suppressed Due to Excessive Length
A A A
15
B B (a) Case 1 (b) Case 2
B (c) Case 3
Fig. 9. Cell types.
to the midpoints of the cube edges (Fig. 10c). The cell generated by the subdivision can be grouped in three main classes: ? Class 1 cell: the cell is cube shaped; ? Class 2 cell: the cell is hexahedral shaped and can be oriented along x, y or z axis; ? Class 3 cell: the cell is octahedral shaped and can be oriented along x, y or z axis;. Figure 9 depicts the three classes just described. Each cell is characterized by a center that correspond to the midpoint of edge AB, longest edge of the cell. The midpoint of AB is always selected as the dividing point and correspond respectively to the center of the cube (Step 1 Fig. 9b), to the center of the cube faces (Step 2 Fig. 9d) and to the midpoint of the cube edges (Step 3 Fig. 9f). In Fig. 10 is shown the entire subdivision process summarized in three steps. The subdivision of Class 1 cells (Fig. 10a) produces cells belonging to Class 2 (Fig. 10b) that subdivided again produce tetrahedra belonging to Class 3 (Fig. 10c). After the subdivision of Class 3 cells the con?guration recursively returns to Step 1. Like Zhou subdivision each type of cell belongs to a di?erent step of the subdivision process and the overall process has a cyclic behavior. The di?erence between the two approaches relies mainly in the cell shape and in the way a cell can be identi?ed. Point worth noting of Pascucci approach is in the fact that to characterize a cell (i.e. type, orientation and re?nement level it belongs to) he just need to know its center. All the approaches presented work under a politics of per-vertex adding process. The subdivision process is always recursive and ends when no more re?nement is needed or if all the vertexes belonging to the mesh has been added. As mentioned before the regularity of the processes allows for an organization of the results in hierarchical data structure suitable of e?cient LOD’s extraction and adaptive traversal.
16
Rita Borgo, Paolo Cignoni, and Roberto Scopigno
(a) Starting step
(b) Step 1
(c) Step 2
(d) Step 3
(e) Step 4
(g) Step 6
Step 1
(f) Step 5
Fig. 10. Cube subdivision by Pascucci et al.
Title Suppressed Due to Excessive Length
17
4 Comparative Evaluation
A comparative evaluation of the two techniques is not straightforward. Both approaches present strength and weakness on di?erent sides. To perform an e?ective evaluation we have then identi?ed six main point of interest that visualization techniques are required to satisfy in the best possible way. – adaptivity: with adaptivity we indicate how well each approach is able to produce di?erent resolution representation of the initial model, which either can be: (a) give a good approximation of the dataset shape or (b) give a good representation of the ?eld distribution encoded in the initial pointset. Irregular techniques allow for better adaptivity due mainly to the freedom in the choice of the best suitable pattern for re?nement purpose. In passing form high resolution to low resolution, and vice-versa, the area on which irregular techniques needs to operate the update can be restricted to a quite small number of tetrahedra. Regular techniques instead need to propagate the change to a wider area because in most cases those type of techniques allows for at most one level of di?erence in the re?nement of adjacent cells. – ?exibility: with ?exibility we indicate to what extent each method allows to change the parameters used to de?ne the simpli?cation criterion. This is a critical feature for what concern the visualization of datasets with multiple ?eld values like, for example dataset coming from ?uidodynamic simulations, in which more than parameters (pressure, temperature) in?uences the error-based re?nement. Regular techniques suit better the possibility of changing re?nement constraints at dynamic level. The regular pattern on which relies the overall re?nement hierarchy allows if needed for fast update of the hierarchy itself. In irregular techniques instead the construction of the re?nement “patterns” is error-driven so a changing in the re?nement parameters, at dynamic time, requires to start the all re?nement process from scratch. – data access: with data access we indicate how well data can be organized in memory to allow for e?cient data access. Irregular techniques require random access to the data on disk, on the contrary regular techniques can be easily organized in memory in such a way that guarantees locality in memory access especially during re?nement actions known only at runtime. – space complexity: with space complexity we indicate how well data can be organized in memory to allow for e?cient storing. The adoption of a regular schema obviously allows for an e?cient representation of the schema itself. With regular techniques it is possible to avoid an explicit representation of the topology of the mesh or piece of mesh under re?nement and often also of the . Irregular techniques requires instead an explicit representation also of the interdependency relationships that exists between the elements that make up the multiresolution structure.
18
Rita Borgo, Paolo Cignoni, and Roberto Scopigno
Regular Irregular
Adaptivity Flexibility Data Access Space Complexity bad bad good good good good bad bad
Two more point of analysis worth to be mentioned would be the implementation easiness and adaptability to di?erent type of datasets. Referring to easiness of implementation it would seem that implementation and debugging of techniques based on regular approaches should be preferable nevertheless reality shows to be di?erent. As the depth of the re?nement augment, even on dataset Mega sized, the structural complexity of the re?nement structure grows exponentially both for the regular and irregular case making quite hard any debugging attempt in both of the approaches. For what concern adaptability irregular techniques show to be more ?exible. Regular techniques in fact performs well only on regular datasets and result to be di?cult to generalize to irregular ones. Re-sampling an irregular dataset on a regular grid is not convenient for several reason. It is common that the ratio between the sizes of the smallest and the largest cell in such a dataset is about 1 : 10, 000, and large cell may lay in any position across the domain. This would require an adaptive re-sampling strategy that often makes it hard using standard techniques peculiar to the regular case. Second most irregular datasets have non-convex domains and the regular grid should properly contain the original data domain. In this case the ?eld has good chances to have a sharp discontinuity cross the boundary of such domain since it is unknown outside the boundary.
5 Conclusions
In this work we have addressed the problem of e?ciently managing large volume datasets using multiresolution structure based on regular and irregular techniques. We have presented the main contribution currently present in literature and analyzed strength and weakness of both the approaches. We have seen how irregular techniques, at expense of a more complex structure, suit better approximation and re?nement of any kind of dataset while regular techniques mainly allows for e?cient re?nement of regular datasets. On their side regular techniques deals better with all the issues arising from the necessity of working in out-of-core. The adoption of regular subdivision pattern allows for e?cient data organization to improve memory occupancy and e?cient access policy to secondary memory properties of prime importance in the visualization of very large datasets. Visualization results from both techniques are considerable even if irregular techniques perform better on adaptive re?nement while regular techniques deals better with change of critical re?nement constraint. Both approaches have been resulted valid for speci?c kind of problems it is still not possible to formulate a unique judgement of quality that could prefer one approach against the other neither this has ever been primary focus of this work.
Title Suppressed Due to Excessive Length
19
References
1. A. Ciampalini, P. Cignoni, C. Montani, and R. Scopigno, Multiresolution decimation based on global error, The Visual Computer 13 (1997), no. 5, 228–246. 2. P. Cignoni, D. Costanza, C. Montani, C. Rocchini, and R. Scopigno, Simpli?cation of tetrahedral volume with accurate error evaluation, Proceedings IEEE Visualization’00, IEEE Press, 2000, pp. 85–92. 3. P. Cignoni, L. De Floriani, C. Montani, E. Puppo, and R. Scopigno, Multiresolution modeling and rendering of volume data based on simplicial complexes, Proceedings of 1994 Symposium on Volume Visualization, ACM Press, October 17-18 1994, pp. 19–26. 4. P. Cignoni, Paola Magillo, Leila De Floriani, Enrico Puppo, and R.Scopigno, Memory-e?cient selective re?nement on unstructured tetrahedral meshes for volume visualization, IEEE Transactions on Visualization and Computer Graphics 8 (2002), no. 3, To appear. 5. P. Cignoni, C. Montani, E. Puppo, and R. Scopigno, Multiresolution modeling and visualization of volume data, IEEE Transactions on Visualization and Computer Graphics 3 (1997), no. 4, 352–369. 6. L. De Floriani, P. Magillo, and E. Puppo, Building and traversing a surface at variable resolution, Proceedings IEEE Visualization 97 (Phoenix, AZ (USA)), October 1997, pp. 103–110. 7. L. De Floriani and E. Puppo, Hierarchical triangulation for multiresolution surface description, ACM Transactions on Graphics 14 (1995), no. 4, 363–411. 8. L. De Floriani, E. Puppo, and P. Magillo, A formal approach to multiresolution modeling, Theory and Practice of Geometric Modeling (R. Klein, W. Stra?er, and R. Rau, eds.), Springer-Velrag, 1997, (to appear). 9. M. Duchaineau, Murray Wolinshy, David E. Sigeti, Mark C. Miller, Charles Aldrich, and Mark B. Mineev-Weinstein, ROAMing terrain: Real-time optimally adapting meshes, Proceedings of the 8th Annual IEEE Conference on Visualization (VISU-97) (Los Alamitos) (Roni Yagel and Hans Hagen, eds.), IEEE Computer Society Press, October 19–24 1997, pp. 81–88. 10. L. De Floriani, E. Puppo, and R. Scopigno, Level-of-detail in surface and volume modeling, IEEE Visualization ’98, 1998, Tutorial Notes #6. 11. R.J. Fowler and J.J. Little, Automatic extraction of irregular network digital terrain models, ACM Computer Graphics (Siggraph ’79 Proc.) 13 (1979), no. 3, 199–207. 12. H. Hoppe, Progressive meshes, Proceedings of SIGGRAPH ‘96 (1996), 99–108. 13. V. Pascucci, Slow growing subdivision (sgs in any dimension: towards removing the curse of dimensionality, (2002). 14. V. Pascucci and C. L. Bajaj, Time critical isosurface re?nement and smoothing, Proceedings of the 2000 IEEE Symposium on Volume visualization 2000 (T. Ertl, B. Hamann, and A. Varshney, eds.), IEEE Computer Society Technical Committe on Computer Graphics, 2000. 15. A. Plaza and G.F. Carey, About local re?nement of tetrahedral grids based on local bisection, 5th International Meshing Roundtable (1996), 123–136. 16. J. Popovic and H. Hoppe, Progressive simplicial complexes, ACM Computer Graphics Proc., Annual Conference Series, (SIGGRAPH 97), 1997, pp. 217– 224.
20
Rita Borgo, Paolo Cignoni, and Roberto Scopigno
17. E. Puppo, Variable resolution terrain surfaces, Proceedings Eight Canadian Conference on Computational Geometry, Ottawa, Canada, August 12-15 1996, pp. 202–210. 18. Mar′ ?a-Cecilia Rivara, Mesh re?nement processes based on the generalized bisection of simplices, SIAM Journal on Numerical Analysis 21 (1984), no. 3, 604–613. MR 85i:65159 19. J. Ruppert and R. Seidel, On the di?culty of triangulating three-dimensional non-convex polyhedra, Discrete Comput. Geom. 7 (1992), 227–253. 20. W.J. Schroeder, J.A. Zarge, and W.E. Lorensen, Decimation of triangle meshes, ACM Computer Graphics (SIGGRAPH 92 Proceedings) (Edwin E. Catmull, ed.), vol. 26, July 1992, pp. 65–70. 21. O.G. Staadt and M.H. Gross, Progressive tetrahedralizations, Proceedings IEEE Visualization ’98, IEEE, 1998, pp. 397–402 (en). 22. I.J. Trotts, B. Hamann, and K.I. Joy, Simpli?cation of tetrahedral meshes with error bounds, IEEE Transactions on Visualization and Computer Graphics 5 (1999), no. 3, 224–237. 23. Yong Zhou, Baoquan Chen, and A. Kaufman, Multiresolution tetrahedral framework for visualizing regular volume data, IEEE Visualization ’97 (VIS ’97) (Washington - Brussels - Tokyo), IEEE, October 1997, pp. 135–142.
京剧行当说课稿 赤峰红山中学白泳鑫尊敬的各位领导、老师： 大家好！ 我今天说课的题目是《京剧行当》，请各位老师批评指正。 首先是教材分析，京剧行当是人音版教材八年级下册京腔昆韵中的拓展延伸课。京剧是中国的国粹，是我国的艺术瑰宝。在我国众多戏曲中，京剧最具有民族性、群众性，代表着中华戏曲文化主流。京剧艺术是博大精深的，其中蕴含的文化不胜枚举。凭借一节课时无法把整体的京剧艺术介绍全面，所以只选出京剧所涉及的一个小点来进行讲解。我所选择的这个点是京剧的行当介绍，因为京剧的行当实践性较强，可以让学生参与体会，进而了解这其中的艺术。 第二、学情特点，我所面对的学生是初中阶段的学生，这一阶段的学生有很强的求知欲，模仿能力较强。让学生通过实践可以更好的理解京剧的行当。 第三、教学理念，本节课是以模仿实践为主的欣赏课 第四、教学目标， 1.学习了解京剧行当，再现一些经典的京剧曲目。 2.通过对京剧行当知识的了解与曲目场景的再现，创设出师生共同参 与的体验活动。 3.通过学习京剧，让学生了解京剧的博大精深，并热爱国粹，喜爱中 国文化。 第五、教学重点难点 教学重点是让学生了解京剧行当，体会京剧的魅力。教学难点是通过体验京剧行当，对京剧功能的再现（实践、体验）。 第六、教学过程 一、导入过程，首先给学生表演一段音配像京胡名曲《夜深沉》，让学生听旋律，观察画面，引出京剧内容，进而往主题上相靠。 二、新课教学 1、介绍京剧定义 首先我会给学生介绍京剧的定义，京剧是中国的国粹，它是由汉调和徽调结合，并吸收了秦腔、昆曲等地方戏曲的精华而成，至今已有两百多年的历史了，表演形式分为唱、念、做、打四种艺术形式，京剧剧目丰富、行当齐全，已被联合国教科文组织列入“人类口头与非物质文化遗产名录”。 2、介绍京剧四大行当 在学生有了一定的了解之后开始详细的介绍京剧的四大行的。 一）介绍“生”角定义，除花脸和丑角以外男性正面角色的统称。性别角色男，分类分为文老生、文小生、红生、武生。代表曲目有《三岔口》。然后欣赏《三岔口》的经典片段。 询问学生视频片段中有哪些伴奏乐器？鼓、钹。让学生体会京剧中伴奏乐器速度的作用，然后带领学生实践，之后练习亮相动作，配合鼓的节奏咚咚嚓一起做出，完成实践环节。 二）介绍“旦”角定义，女性正面角色的统称。角色性别为女，行当分类分为青衣，代表人物白素贞；花旦，代表人物春草；刀马旦，代表人物穆桂英；老旦，代表人物佘太君。这些都是按照年龄、身份、性格表演特点而分的。然后欣赏旦角的经典片段《卖水》，欣赏之后让学生参与剧中经典场景。让学生体会“旦”
大学学生简单自我介绍 大学学生简单自我介绍 第一篇： 大学学生简单自我介绍 你好!我名字叫haoord，我22岁。来自xx，一个美丽的文化城市，现在就读于xx大学。一般来说，我是一个勤奋的学生，尤其是做我感兴趣的东西。我会尽我最大的努力，无论是多么困难的事情，都要完成它。在我读大二的时候，我发现网站设计很有趣，所以我很努力的学习。通过两个月的时间，为自己制作了一个网页。我是我们班第一个拥有自己主页的人。 此外，我是一个很有毅力的人。我坚持每天跑步，不管什么天气。在我的业余时间，我喜欢篮球，网球和中国象棋。英语也是我的最爱。我经常去英语角练习我的英语口语，写作文，以改善我的写作能力。但我知道我的英语不够好，我会继续学习。谢谢! 第二篇： 简单的英文自我介绍 mature，dnami and honest。思想成熟、精明能干、为人诚实。 exellent abilit of sstematial management。有极强的系统管理能力。 be elegant and ith nie personalit。举止优雅、个人性格好。 ith good managerial skills and organizational apabilities。有良好的管理艺术和组织能力。
the main qualities required are preparedness to ork hard, abilit to learn, ambition and good health。主要必备素质是吃苦耐劳精神好、学习能力优、事业心强和身体棒。 ith job searhes taking longer, there are more jobseekers bumping up against their severane deadline 。随着找工作花的时间长了，越来越多的求职者发现已逼近他们找工作的明确限期，或者 仅仅是他们心理上的限期，即主观上告诉自己现在本该已找到工作 的。 a popular question that i hear often is ho long ou should hold out for that dream jo b before moving to plan b。我常听到的一个问题就是： 人们应该为梦想中的工作坚持多久才退而执行另外的计划呢? the anser is different for everone:答案因人而异，下面一些可能的影响因素： hat is our time urgen? have ou run the atual numbers on our ash ushion runs out? the shorter the time ou have, the ider our net has to be. even if our dream pan industr funtional area is hiring, hen ou have a short timeframe, ou need to have multiple searhes ell undera。你最紧迫的事是什么?你有没有计算过你目前的储蓄可以支撑多少天?时间越急，你撒的网就必须越广。即使你心仪的公司、行业、领域正在招人，考虑到时间紧促，你也应该同时多处留意。 hat is our risk appetite? some people are energized b putting all their fous on their passion. some people ould be
生、旦、净、丑是京剧的四大行当。 1、生行 生行是扮演男性角色的行当，在京剧中的地位非常重要。生行包括老生、小生、武生、红生、娃娃生等几个门类。除红生和勾脸武生以外，生行一般都是素脸，内行术语称为“俊扮”，即扮相都是洁净俊美的。 ①老生又称“须生”、“正生”、“胡子生”。一般都是富有正义感的男性中年或老年人物。人物形象以挂“黑髯口”（黑胡子）为主。按照表演艺术特点的不同，老生又分为安工老生、衰派老生和靠把老生。安工老生，又称“唱切老生”、“王帽生”、大部扮演帝王、书生一类人物，以唱功为主，在舞台上安详稳重，动作较少，故称“安工”，如《上天台》中的汉光武帝刘秀、《捉放曹》中的陈宫等。衰派老生，又称“做功老生”。“做工老生”，大都扮演衰老或精神状态衰颓的人物，以做功为主，故称“衰派”，如《四进士》中的宋土杰、《卖马》中的秦琼、《坐楼杀惜》中的宋江等。靠把老生，大都扮演武将一类的人物，由于扎靠（戴盔披甲）、使用刀枪把子（剧中人使用的兵器）而得名，加《定军山》中的老将黄忠、《战太平》中的花云、《镇潭州》中的岳飞等。无论是哪类老生，都是以唱为主，全用本嗓。安工老生要唱的悠扬婉转，衰派老生要唱的悲愤颓唐，靠把老主要唱的激昂慷慨。另外，京剧史上有一些老生演员，文戏、武戏都擅长，唱功戏、做功戏、靠把戏都能演，后来就把这种戏路宽的老生演员称为“文武老生”，如程长庚、谭鑫培等。 ②小生主要扮演青少年男子，化妆不戴胡须，扮相清秀、英俊。小生演唱用尖音假嗓，念白兼用真假嗓。小生使用的假嗓与旦角不同，小生的唱法应该刚、劲、宽、亮，听起来声音清跪但不柔媚，刚健但不粗野。根据人物性格、身分的不同特点，小生又分为袍带小生、扇子生、翎子生、穷生和武小生。袍带小生，又称“纱帽小生”，扮演做官的青年男子，头戴纱帽是其主要标志，这些角色大部分是文人，如《玉堂春》中的王金龙、《奇双会》中的赵宠、《陈三两爬堂》的陈魁等。扇子生，多扮演年青的书生、风流儒雅的公子，手拿扇子，头戴文生巾，身穿褶子，所以又称“褶子生。”扇子是帮助角色表现风流潇洒、文质彬彬的一种特殊道具。如《拾玉镯》中的傅朋、《西厢记》中的张君瑞等。翎子生，又称“雉尾生”，头插翎子（雉尾）是其主要标志，大都扮演武将或文武兼备的人物。演翎子生要有武功，凭工架、舞蹈，用翎子耍出许多舞蹈动作。如《群英会》中的周瑜、《吕布与貂蝉》中的吕布、《穆柯寨》中的杨宗保等等。穷生，大部扮演落魄不第的文人、书生，“表演上特别注重做功，”以表现人物的酸腐气为主，习惯于把鞋后帮踩倒在脚下，以示其潦倒之状，故又称“鞋皮生”，身穿富贵衣是其主要标志。富贵衣，是补缀有许多五颜六色的补丁的青褶子，意思是说这些人现在虽然很穷，衣着褴楼，但将来仍要腾达，故有“富贵衣”之名。如《棒打薄情郎》中的莫稽、《连升店》中的王明芳等。武小生，大多扮演年青英武的人物，表演著重武功，也兼重唱功、念白和做功。从武打功夫上看，武小生与武生差不多，但是唱与念全用小生方法。如《八大锤》中的陆文龙、《借赵云》中的赵云、《石秀探庄》中的石秀等。 ③武生扮演擅长武艺的青壮年男子，分为长靠武生和短打武生两类，俗称“墩子武生”、“撇子武生”，“墩子”是指穿厚底靴子，“撇子”是指穿薄底靴子。长靠武生扎大靠，武打、工架并重，如《长坂坡》中的赵云、《战冀州》中的马超、《挑滑车》中的高宠等。短打武生身穿紧身短装。偏重武打特技，如《三岔口》中的任堂惠、《十字坡》中的武松、《夜奔》中的林冲等。扮演中老年英代人物的称武老生，如《百凉楼》中的吴祯、《剑峰山》中的邱成等。武生还兼演部分勾脸戏（武净戏），如《铁笼山》中的姜维、《拿高登》中的高登等。猴戏中的孙悟空一般也由武生扮演。 ④红生指勾红脸的老生，主要扮演关羽、赵匡胤等角色，演唱嗓音高亢浑厚，表演具
大学生个人简短自我介绍范文 大学生个人自我介绍范文篇一 各位面试老师好，我叫xxx，是医科大学医学检验系的学生，将毕业并获得学士学位。医学检验是一门对动手能力要求较高的学科，所以，在校期间，除了医学基础课和专业课理论的学习，我们还被安排了大量的实验课，其中临床检验、微生物检验、生化检验、免疫学检验等专业课，他们的实验课比理论课还要多，这些既使我们对理论有了更深的理解和掌握，更重要的是使我们有了熟练的动手操作能力。我在随后的xx市第一中心医院检验科实习期间，很快就得到了代教老师的认可和信任，而肯放开手让我独立操作仪器和独立进行实验，我也就更加积极主动地去做得更好，遇到不明白的问题及时地向老师请教。等到实习结束时已经熟知了检验科的工作流程和具备了大多实验的独立操作能力。 最后我由衷的希望您能够给我一次展现才华的机会，能让我加入贵单位。当然了，刚刚跨出大学校门的我在工作经验和社会阅历方面显得捉襟见肘，但我深信天道酬勤，我愿意从最一点一滴做起，我相信经过加倍努力，并在工作中不断学习，我相信一定能赢得您和大家的认可。 大学生个人自我介绍范文篇二 我是***学院***专业的学生。 大学四年中，我各方面的能力都得到了发展，可以说，经过大学四年的学习，我已经具备了适应社会工作的能力。这学期即将画上了句号，就是毕业了。回首往事，至少可以自信地说一声“我没有虚度”。有必要对这四年做个自我评定。 在思想上，我要求上进，一直以乐于助人为已任，多次参加青年志愿者活动。尊敬师长，团结同学，为自己的学习和生活创造了良好的环境。 在学习上，我刻苦努力，孜孜不倦，争取着大学那美好的时光去学习。大学四年，不光使我学到了许多知识，也使我懂得了学习的方法。正是利用这种方法，在除学校开设的课程外，我还自学了网络数据库、网页制作、平面设计等知识，很好地充实了自己的业余生活，并为自己的将来打下良好的基础。到目前为止，我已掌握了本专业的基础知识和有关网络的基本知识。除此之外，对计算机的爱好让我对计算机有一定的了解，并具有一定的编程能力。工作方面，我参与了
基于大数据的能力开放平台解决方案 1 摘要 关键字：大数据经分统一调度能力开放 运营商经过多年的系统建设和演进，内部系统间存在一些壁垒，通过在运营商的各个内部系统,如经分、VGOP、大数据平台、集团集市等中构建基于ESB 的能力开放平台,解决了系统间调度、封闭式开发、数据孤岛等系统问题，使得运营商营销能力和效率大大提高。 2 问题分析 2.1 背景分析 随着市场发展，传统的开发模式已经无法满足业务开发敏捷性的要求。2014 年以来，某省运营商经营分析需求量激增，开发时限要求缩短，业务迭代优化需求频繁，原有的“工单-开发”模式平均开发周期为4.5 天，支撑负荷已达到极限。能力开放使业务人员可以更便捷的接触和使用到数据，释放业务部门的开发能力。 由于历史原因，业务支撑系统存在经分、VGOP、大数据平台、集团集市等多套独立的运维系统，缺乏统一的运维管理，造成系统与系统之间的数据交付复杂，无法最大化 的利用系统资源。统一调度的出现能够充分整合现有调度系统，减少运维工作量，提升维护质量。 驱动力一：程序调度管理混乱，系统资源使用不充分
经分、大数据平台、VGOP、集团集市平台各自拥有独立的调度管理，平台内程序基本是串行执行，以经分日处理为例，每日运行时间为20 个 小时，已经严重影响到了指标的汇总展示。 驱动力二：传统开发模式响应慢，不能满足敏捷开发需求 大数据平台已成为一个数据宝库，已有趋势表明，只依赖集成商与业 务支撑人员的传统开发模式已经无法快速响应业务部门需求，提升数据价值。 驱动力三：大数据平台丰富了经分的数据源，业务部门急待数据开放 某省运营商建立了面向企业内部所有部门的大数据平台，大数据平台 整合了接入B域、O 域、互联网域数据，近100 余个数据接口，共计820T 的数据逐步投入生产。大数据平台增强了传统经分的数据处理的能力，成为公司重要的资产，但是传统经分数据仓库的用户主要面向业支内部人员，限制了数据的使用人员范围和数据的使用频度，已经无法满足公司日益发展的业务需求，数据的开放迫在眉睫。 2.2 问题详解 基于背景情况分析，我们认为主要问题有三个： 1、缺乏统一的调度管理，维护效率低下 目前经分系统的日处理一般是使用SHELL 脚本开发的，按照串行调度的思路执行。进行能力开放后，目前的系统架构无法满足开发者提交的大量程序执行调度的运维需求。如果采用统一调度的设计思路则基于任务的数据表依赖进行任务解耦及调度，将大大简化调度配置工作和提高系统的
京剧四大行当
京剧四大行当——生、旦、净、丑 行当是我国传统戏曲一种重要的表演艺术手段，各地方戏曲剧种都有行当的划分。 京剧原分十大行当：一末二净三生四旦五丑六外七小八贴九夫十杂；后概括为五大行当：生旦净末丑；最后归并为“生旦净丑”四大行当。 除这四行外，尚有几类次要的角色，如武戏中的配角“武行”和“觔斗行”，以及跑龙套的“流行”，所以也有七行之说，即“生旦净丑，武流觔斗”。其他地方剧种的行当和京剧大同小异，现我们就京剧的行当作一简单的介绍—— 1、生行 生行，指戏曲剧目中的男性形象，以面部化妆为俊扮为其特点。根据其年龄、身份的不同可以分为老生、小生、武生等不同的种类。 老生：指生行中的中年或老年形象，以戴胡须（即髯口）为其特点。又根据其不同的表演特点，分为唱功老生、做功老生和靠把老生三类。 小生：指生行中的年轻人形象，根据不同的表演特点，可分为扇子生、纱帽生、雉尾生、穷生、武小生等类别。（《群英会》的周瑜是雉尾小生） 武生：指生行中身具武艺的男性形象，又有长靠武生（表现大将）和短打武生（表现绿林英雄）之分。前者扎大靠，重工架；后者穿紧身衣服，重翻打。 （如《长坂坡》之赵云) 红生：有时将其归入武生行，指勾红脸的身具武艺的男性形象，最典型的是 关羽和赵匡胤。 2、旦行 旦行，指戏曲中的女性形象，可分为青衣、花旦、刀马旦、武旦、老旦、彩旦等类别。 青衣：又叫“正旦”，多表现那些端庄稳重的中青年妇女，以唱功见长，如《铡美案》中的秦香莲，《二进宫》中的李艳妃等。 花旦：多表现那些年轻活泼俏俐的小家碧玉或丫鬟，以做功和念白见长，如《西厢记》中的红娘，《拾玉镯》中的孙玉姣等。 武旦：表现那些身俱武艺的江湖女子或神怪精灵，多穿紧身衣服，表演上重翻打，如《白蛇传》中的青蛇、《十字坡》的孙二娘等。
大学生的简单自我介绍 大学生的简单自我介绍1 各位老师、同学，大家好，我叫宋轩，1988年6月6日出生，20岁，来自山东，齐鲁大地给了我直爽的性格，但又不失稳重，之后不远千里来到南京这个城市求学。 来到大学学习的事实和我的理想有很大的出入，难免有些郁闷，但在一段时刻后，我认清了事实，很看互联网专业。”21世纪是网联网的世纪。”这句话一点都不假，随着互联网的展，它为21世纪插上了腾飞的翅膀。之后在不断的培养兴趣过程中，我开始对互联网产生兴趣，今后的四年，我将在不断的学习进步中度过的。 “十年磨砺锋利出，宝剑只待君来识”。再苦再累，我都愿意一试，“吃得苦中苦，方为人上人”，在以后的学习生活中，我必须会是一位尽自己的发奋、过一个充实而又好处的大专生活。 大学生的简单自我介绍2 尊敬的领导、敬爱的老师、亲爱的同学们、朋友们：大家好! 踏着清爽的秋风，吻着醉人的菊香，在这丰收的季节，我从千里之外的辽东来到了素有天堂美誉之称的杭州，跨进了梦寐以求的浙江商业学院的大门，开始了我的大学之梦的旅程。 当我踏进这所高等学府的大门时，看到老师们热情的手
臂，温暖的笑容，听到新同学们的欢歌笑语和亲如弟兄姐妹般的问候，让我周身的血液开始沸腾，心跳也开始加速。这是为什么呢?正因那里就是我人生的一个最最重要的转折点，一个让我圆梦的地方!读大学是我们每一个年轻人孜孜以求的梦想，这个梦从我的儿时就开始了。那时候虽然我还不懂得什么是理想、什么是奉献、什么是职责，可我羡慕科学家、羡慕作家、羡慕宇航员、羡慕一切有作为的精英干将。渴望自己将来也能够像他们一样，成为他们其中的一员，为伟大祖国的建设事业做出自己的贡献。历经十二年的中小学的读书过程，让我逐渐明白了一个道理，那就是：要想成功，一要有聪睿的头脑，优良的道德品质，二要有渊博的知识，丰富的学识。三是要有健康的体魄，宽旷的胸怀。最后就是要有上天赐予的机遇。而机遇是给有准备的人准备的，这个准备就是第一、第二和第三的资料，那么大学就是我们做这个准备的最好地方。只有做好了第一、第二和第三的准备工作，我们才能在机遇到来之时，从容不迫地伸出我们的双手去迎接他、应对他，而被他所接纳、所包容。我们也就能在自己胜任的舞台上，演好自己的主角。以不负家乡父老的殷殷期望，老师的辛苦栽培。 愚者错失机会，智者善抓住机会，成功者创造机会，机会只是给准备好的人。机遇真是“神奇”，它能给“山穷水尽疑无路”的人带来“柳暗花明又一村”的喜悦;它能让商人散尽千金“还复重来”……。“机遇”说起来很神奇，其实它经常出此刻我们的身边，而智者能发现它、利用它走向
大学简短幽默自我介绍_大一新生开 学有趣的自我介绍 幽默是人类独有的品质、能力和交际方式。大学生用幽默的话语来做自我介绍可以得到同学们的好感。下面是小编为大家整理的大学简短幽默自我介绍，仅供参考。 大学简短幽默自我介绍篇一 本人外表平平，在此就不细说了，大街上随便抓个人都会有和我的相识之处。性格还好，中庸是我的理想，但迫于社会现实所逼，经常做出一些左倾和右倾的事情，于是乎导致我现在没多大成就，除了找到几个我十分玩命的朋友，我感觉自己还是相当孤独的。不过只有高处的人才会感到孤独，与我自身的空虚寂寞——也就是年轻人特有的茫然无关。 我深刻感受到自己的善良，于是我坚定好人有好报，因为我无意中的善行为我赢来了许多朋友，感觉是上天的馈赠，但偶然我也觉得自己很无情，尤其是见到四肢健全、身强力壮的乞丐，我都极力装没看到。作为一男生，我又非常要脸皮，非常要面子，而且有点过火了，可能我的脸皮特别厚，所以也特别要面子，不
能让我的脸皮受一点损害。呵呵!差不多了，这个人很简单，又不是大人物，没太多讲的。 大学简短幽默自我介绍篇二 大家好，秋天来了，我也来了.我叫某某.很高兴能站在这里跟大家做个自我介绍.我很喜欢笑，妈说我出生的时候是笑着生出来的.为这事中科院的找我好几回，非要给我做个全面检查.我没同意，我说除非你们的人当中有谁比我的眼睫毛张得好看，嘿嘿，大家注意到没有.我的眼睫毛很好看哦，不信? LOOK!如果现在你看不清楚，没有关系，下课了找我啊，我很愿意交朋友，因为我十分开朗，交到我这个朋友，免费让你看个够，赶快行动吧!谢谢! 大学简短幽默自我介绍篇三 大家好我叫XXX 我是一个超级活泼的人噢~` 不要被我看似文静或者内敛的外面给欺骗了噢~我的胜任座右铭是：态度决定一切。因为我觉得一个人到底会不会成功，态度才是最关键的因素。我喜欢看什么什么书~ 曾经记得那里有这么一句话~《要牛逼一点的》我觉得它是能给我的人生一点启示的话所以拿出来与大家分享，希望大家也能从中得到一点帮助。
某大型企业数据平台整体解决方案
目录 1项目概述 (15) 1.1建设背景 (15) 1.1.1集团已有基础 (15) 1.1.2痛点及需提升的能力 (15) 1.1.3大数据趋势 (16) 1.2建设目标 (16) 1.2.1总体目标 (16) 1.2.2分阶段建设目标 (17) 1.3与相关系统的关系 (18) 1.3.1数据分析综合服务平台 (18) 1.3.2量收系统 (19) 1.3.3金融大数据平台 (20) 1.3.4各生产系统 (20) 1.3.5CRM (20) 1.4公司介绍和优势特点 (20) 1.4.1IDEADATA (20) 1.4.2TRANSWARP (22) 1.4.3我们的优势 (24) 2业务需求分析 (27) 2.1总体需求 (27)
2.2.1数据采集 (29) 2.2.2数据交换 (29) 2.2.3数据存储与管理 (29) 2.2.4数据加工清洗 (30) 2.2.5数据查询计算 (31) 2.3数据管控 (32) 2.4数据分析与挖掘 (32) 2.5数据展现 (33) 2.6量收系统功能迁移 (34) 3系统架构设计 (35) 3.1总体设计目标 (35) 3.2总体设计原则 (35) 3.3案例分析建议 (37) 3.3.1中国联通大数据平台 (37) 3.3.2恒丰银行大数据平台 (49) 3.3.3华通CDN运营商海量日志采集分析系统 (63) 3.3.4案例总结 (69) 3.4系统总体架构设计 (70) 3.4.1总体技术框架 (70) 3.4.2系统总体逻辑结构 (74)
3.4.4系统接口设计 (83) 3.4.5系统网络结构 (88) 4系统功能设计 (91) 4.1概述 (91) 4.2平台管理功能 (92) 4.2.1多应用管理 (92) 4.2.2多租户管理 (96) 4.2.3统一运维监控 (97) 4.2.4作业调度管理 (117) 4.3数据管理 (119) 4.3.1数据管理框架 (119) 4.3.2数据采集 (122) 4.3.3数据交换 (125) 4.3.4数据存储与管理 (127) 4.3.5数据加工清洗 (149) 4.3.6数据计算 (150) 4.3.7数据查询 (170) 4.4数据管控 (193) 4.4.1主数据管理 (193) 4.4.2元数据管理技术 (195)
简单的大学生自我介绍范文 各个大学的大一新生们，已经开始露露开学了，开学就是面临着新的生活了，在进入校园后免不了和大家分班级，每一个人都要上台在全班同学面前做自我介绍，这个时候大家一定要明白有很东西是不可以说的，否则会让自己交不到朋友，甚至会让同学们鄙视，下面就为大家分享四个不能介绍的点，想要了解的朋友们，赶紧与来看看吧! 一、爱好 ___喜欢吃什么和玩什么 相信有不少的同学，在介绍自己时都会说自己的爱好，当自己向大家说喜欢吃什么和玩什么时，肯定是不合适的，大家要明白跟初高中时的介绍是不一样的，大学的四年生活与自己关系最密切的还是室友，除了寝室的几个人，与其他的同事关系根本不会密切，大家在介绍自己时，就没有必要在台上说的没完没了，这样做只会浪费大家的时间，给同学们留下不好的印象;因此，大一新生在介绍自己时，千万不要说一些无关紧要的事情，最好控制在3-5分钟最好。 二、关于爸妈的事情最好不要说
有很多大一新生，在介绍自己时，都会介绍自己的爸妈，以及自己的成长环境;大家在介绍自己时要明白，那些愿意介绍爸妈工作的人，一般家里的条件都不是太差的，不管自己的爸妈是企业家，还是干部工作者，只有是强调了爸妈的成就，这些话的潜在意思就是告诉大家，你的家庭条件好，财富底子厚，会给大家一种不好的感觉，甚至会起到相反的作用，试想那些家庭普通的同事，怎么敢靠近你这个富人子弟，相信大部分人都不愿意和喜欢炫耀的人相处;因此，大一新生在介绍自己时，不要介绍关于爸妈的背景。 三、平时成绩好高考发挥失利的话不要说 也有些同学，上台后介绍自己时，会说一些关于自己高中成绩的事情，自己平时的成绩有多好，自己高考发挥失利，只考了多少多少分，造成了自己和大家成为同班同学，当你向大家说这些话时，大家就会认为你是在看不起人家，看不起现在的学校和班级，相信大部分同学都会对你产生反感;大家要明白不管自己以前怎么样，在进入大学后就是新的开始，大家又重新站在统一起跑线上了，明白这个道理大家才会珍惜和你之间的关系;因此，大家不要随便介绍，平时成绩好高考发挥失利的事情。 四、自己的感情状态不要说
Solution 解决方案 大数据平台安全解决方案 防止数据窃取和泄露确保数据合规使用避免数据孤岛产生 方案价值 大数据平台安全解决方案为大数据平台提供完善的数据安全 防护体系，保护核心数据资产不受侵害，同时保障平台的大数据能被安全合规的共享和使用。 数据安全防护体系以至安盾?智能安全平台为核心进行建设。智能安全平台支持三权分立、安全分区、数据流转、报警预警和审计追溯等五种安全策略，以及嵌入式防火墙、访问控制、安全接入协议等三道安全防线，保证安全体系在系统安 全接入、安全运维、数据流转、数据使用、数据导出脱敏、用户管理、用户行为审计追溯等方面的建设，保障大数据平台安全高效运行。 智能安全平台提供安全云桌面，保证数据不落地的访问方式， 并可根据需求提供高性能计算资源和图形处理资源，并支持“N+M”高可靠性架构，保证云桌面的稳定运行，为平台用户提供安全高效的数据使用环境。 提供数据不落地的访问方式以及完善的文档审批和流转功能 提供五种安全策略和三道安全防线提供严格的用户权限管理和强大的用户行为审计和追溯功能 提供高性能、高可靠稳定运行的大数据使用环境 方案亮点 如欲了解有关志翔科技至安盾? ZS-ISP、至明? ZS-ISA安全探针产品的更多信息，请联系您的志翔科技销售代表，或访问官方网站：https://www.doczj.com/doc/5a1267658.html, 更多信息 志翔科技是国内创新型的大数据安全企业，致力于为政企客户提供核心数据保护和业务风险管控两个方向的产品及服务。志翔科技打破传统固定访问边界，以数据为新的安全中心，为企业构筑兼具事前感知、发现，事中阻断，事后溯源，并不断分析与迭代的安全闭环，解决云计算时代的“大安全”挑战。志翔科技是2017年IDC中国大数据安全创新者，2018年安全牛中国网络安全50强企业。2019年，志翔云安全产品入选Gartner《云工作负载保护平台市场指南》。 关于志翔科技 北京志翔科技股份有限公司https://www.doczj.com/doc/5a1267658.html, 电话： 010- 82319123邮箱：contact@https://www.doczj.com/doc/5a1267658.html, 北京市海淀区学院路35号世宁大厦1101 邮编：100191 扫码关注志翔
教学科研大数据平台 解决方案
目录 1.概述 (3) 1.1.背景 (3) 1.2.建设目标 (3) 1.3.建设的步骤和方法 (3) 2.教学科研大数据平台概要 (4) 2.1.架构设计 (4) 2.2.教学科研大数据平台优势 (6) 2.2.1.应用优势 (6) 2.2.2.未来发展优势 (8) 3.教学科研大数据平台设计 (8) 3.1.大数据资源池 (9) 3.1.1.cProc云计算 (9) 3.1.1.1.cProc云计算概述 (9) 3.1.1.2.数据立方 (10) 3.1.1.3.混合存储策略 (15) 3.1.1.4.云计算核心技术 (15) 3.1.1.4.1.数据处理集群的可靠性与负载均衡技术 (15) 3.1.1.4.2.计算与存储集群的可靠性与负载均衡 (19) 3.1.1.4.3.计算与存储集群的负载均衡处理 (21) 3.1.1.4.4.分布式文件系统的可靠性设计 (23) 3.1.1.4.5.分布式数据立方可靠性设计 (23) 3.1.1.4.6.分布式并行计算可靠性设计 (25) 3.1.1.4.7.查询统计计算可靠性鱼负载均衡设计 (25) 3.1.1.4.8.数据分析与数据挖掘 (27) 3.1.1.4.9.cProc云计算优势 (35) 3.1.2.cStor云存储 (36) 3.1.2.1.cStor云存储介绍 (36) 3.1.2.2.cStor云存储架构 (38) 3.1.2.3.Stor云存储关键技术 (43) 3.1.2.4.数据安全诊断技术 (44) 3.1.2.5.cStor云存储优势 (45) 3.2.大数据教学基础平台 (46) 3.2.1.Hadoop架构 (46) 3.2.2.Hadoop关键技术 (47) 3.2.3.Hadoop优势 (51) 3.2.4.Hadoop教学 (51)
农村大数据平台解决方案
时间：2018年9月
1大数据服务基础平台 (1) 2农村大数据资源中心 (2) 2.1涉农信息基础大数据 (2) 2.2农业产业技术数据 (2) 2.3农村生活信息服务数据 (3) 2.4政务应用数据 (3) 3大数据共享平台 (3) 4大数据分析平台 (3) 4.1区域经济分析 (4) 4.2生产智能化大数据平台 (4) 4.3农产品质量安全追溯大数据应用 (5) 4.4农产品产销信息监测预警大数据分析 (5) 5智慧农业云平台 (6) 6大数据精准扶贫 (6) 7农村网络舆情监测平台 (7)
农村大数据平台解决方案 根据《关于实施乡村振兴战略的意见》（中发〔2018〕1号）、《农业部办公厅关于印发〈农业农村大数据试点方案〉的通知》（农办市〔2016〕30号）、《农业部关于印发〈”十三五”全国农业农村信息化发展规划〉的通知》（农市发〔2016〕5号）、《农业部关于推进农业农村大数据发展的实施意见》（农市发〔2015〕6号）和《国务院关于印发促进大数据发展行动纲要的通知》（国发〔2015〕50号）等有关部署文件要求，公司经过大量的调研和论证，集中技术力量研发的一整套针对我国农村农业现状的大数据平台产品体系，包含农村大数据基础服务平台、农村大数据资源中心、大数据共享平台、大数据分析平台、智慧农业云平台、大数据精准扶贫、农村网络舆情监测平台等产品。 1大数据服务基础平台 作为农村大数据平台的核心与基础，集成了大数据平台的多个底层组件，提供分布式存储（HDFS）、分布式计算、协调服务管理、数据仓库SQL服务、NoSQL数据库服务,分布式内存计算，ETL 调度与操作，实时流处理、分布式内存、索引搜索、数据库联邦查询、MPP数据库服务，图数据库和时序数据库等功能和服务。同时支持大数据的分布式机器学习算法比如多重估值算法。 平台基于镇平县农业大数据研究的个性化需求，形成一系列相关公开发布数据的采集机制，将数据采集的相关程序设计并编写完善，部署此套机制在平台上周期运转；为管理人员与数据工程师提供数据的浏览，对数据进行查询、展现和基础统计分析等初步应用，实现农业大数据分析人员的交流平台。 1
老生介绍 老生又叫须生，或胡子生。因为老生都挂胡子。胡子在京剧里的专门名词叫“髯口”。老生除了须生和胡子生，还有一个名词叫正生，表示严肃端庄的意思。老生主要扮演中年以上的男性角色，唱和念白都用本嗓(又叫真声或大嗓)。照规矩，老生基本上都是戴三绺的黑胡子(术语叫作“黑三”)。另外还有灰色的三绺胡子，即花白的三绺胡子，专门名词叫“黪(音惨)三”，或“苍三”。还有白色的三绺胡子，叫作“白三”。还有一种胡子，不分绺，就是整片满口的胡子，术语叫作“满”，如黑满、白满、黪满。 照京剧原来的传统，不论是戴黑满的，或是戴黪三、黪满、白三、白满的，都不属于老生扮演的范围。只有戴黑三的，才能算是真正的老生。其他的角色都由“末”行或“外”行来应工。现在“末”“外”都合并到老生行当里，老生扮演角色的范围大为扩大，也就没有什么区别了。 老生一般分为文武两种，从表演的侧重点来划分，可以分成这么几种：唱工老生，做工老生，武老生。唱工老生也叫作安工老生，以唱为主。为什么叫安工老生呢?我想是因为这种老生既以唱为主，那么他的动作性就比较次要，动作幅度较小，态度比较安闲从容，唱的时候总是比较沉着安稳的，因此叫作安工老生。例如《二进宫》的杨波，《捉放曹》的陈官，《击鼓骂曹》的祢衡，《文昭关》和《鱼肠剑》的伍子胥，《洪羊洞》和《辕门斩子》的杨延昭，《空城计》、《雍凉关》、《战北原》、《脂粉计》、《七星灯》等剧的诸葛亮，《四郎探母》的杨延辉，《碰碑》的杨继业。《桑园寄子》的邓伯道，《让徐州》的陶谦，《逍遥津》的汉献帝等等，都属于唱工戏。还有一种介乎唱工戏和做工戏之间的以念白为主的戏，唱工老生和做工者生都可以兼演，仍如《十道本》的褚遂良，《夜审潘洪》的寇准，《审头刺汤》的陆炳，《审刺客》的闵觉等，都属于这一类。由于大段念白很难掌握，从表现技巧看，比起唱工和做工戏，难度大得多，所以这类戏现在很少演出了。做工老生又叫作哀派老生，是以表演为主的一种行当。衰派老生也有两种解释，一种是说，这类老生专门扮演年老体弱、戴白胡子的角色，所以叫作衰派。例如《清风亭》的张元秀，《扫松下书》的张广才，《三娘教子》的薛保，《九更天》的马义，《徐策跑城》的徐策等等，都屑于这一类。另一种解释说是因为以做工为主，一般都是表演精神上受了刺激，情绪紧张，神态粗犷，动作激烈的角色，这些角色的精神状态近乎颓唐衰朽，所以叫作衰派。例如《坐楼杀惜》的宋江，《打棍出箱》的范仲禹，《失印救火》的白槐，《春秋笔》的驿丞张恩，《战蒲关》的刘忠等等，都属于这一类。我认为衰派老生这个名词并不全面，还是说做工老生，特点就是侧重以表演为主，这样比较科学、严谨。唱工老生和做工老生都属于文老生的范畴。 武考生包括长靠和箭衣(俗称短打)两种。长靠老生又称靠把老生。“靠”是京剧的专门名词，就是古代武将所穿的铠甲。身穿铠甲，在京剧里叫作披靠或扎靠。“把”是“把子”的简称，就是兵器，俗称刀枪把子，也是京剧的专门名词。凡是身披铠甲，手持兵器，擅长武功的老生角色，都叫作靠把老生。例如《定军山》的黄忠，《阳平关》的黄忠，《失街亭》的王平，《战太平》的花云，《震潭州》、《八大锤》等剧中的
大学生简单的自我介绍范文推荐 如何做一个好的自我介绍也是大学生们在写简历时经常碰到的问题。今天就给大家分享几篇大学生简单的自我介绍优秀范文，大家一起随来看看吧。 我叫李添添，1992年6月1日出生，17岁，来自西安，我有着直爽的性格，但又不失稳重，不远千里来到怀化这座城市求学。 来到中专学习的事实和我的理想有很大的出入，难免有些郁闷，但在一段时间后，我认清了事实，很看计算机应用专业。"21世纪是电脑的世纪."这句话一点都不假，随着电脑的展，它为21世纪插上了腾飞的翅膀.后来在不断的培养兴趣过程中，我开始对电脑产生兴趣，今后的三年,我将在不断的学习进步中度过的。 “十年磨砺锋利出，宝剑只待君来识”。再苦再累，我都愿意一试，“吃得苦中苦，方为人上人”，在以后的学习生活中，我一定会是一位尽自己的努力、过一个充实而又意义的中专生活。 大学自我介绍篇2您好!我是XX大学XX学院XX专业毕业生，惠于学校浓厚的学习、创新氛围，熔融其中的我成为了一名复合型人才。大学的四年使我思想和知识,心理都得到了进一步的成长。 从入学以来，我立志要在大学四年里全面发展自己，从适应社会发展的角度提高个人素质。将来真正能在本职工作上做出成绩，为母校争光。在学习上我勤奋严谨，在掌握了本专业知识的基础上，不忘
拓展自己的知识面，对课外知识也有比较广泛的涉猎。我还很重视英语的学习，不断努力扩大词汇量，英语交际能力也有了长足的进步。 同时，为了全面提升个人素质，我积极参加各种活动，这些经历使我认识到团结合作的重要性，也学到了很多社交方面的知识，增加了阅历，相信这对我今后投身社会将起重要作用。现在虽存在着很多艰难和困苦,但我坚信大学生活给我的精神财富能够使我战胜它们。“长风破浪会有时，直挂云帆济沧海”，希望贵公司能给我一个发展的平台，我会好好珍惜它，并全力以赴，为实现自己的人生价值而奋斗，为贵公司的发展贡献力量。 大学自我介绍篇3我是一个性格稳重，心理素质较好的人。做事踏实认真，热情主动，吃苦耐劳，并且敢于承担责任。平时喜欢和他人沟通，善于与人相处，具有良好的团结合作精神。 四年来，我一直怀着对本专业的热爱，认真学习专业知识。学习了电路分析，微机原理，C/C++编程，数字/模拟电子技术，单片机，DSP，电路CAD等等一系列与电子相关的课程;出于对嵌入式开发的爱好,我自学了许多嵌入式方面的知识,比如ARM嵌入式系统，Linux 操作系统,Linux C编程,uc/GUI等。 参加了西南大学创新基金项目，并担任主要队员,培养了团队合作意识，学到了一定的开发经验，并了解到了很多先进技术。作为一名电子专业的大学应届毕业生，我所拥有的是年轻和知识。年轻虽然缺少经验，但是年轻也意味着热情和活力，我自信能凭自己的能力和学识在毕业以后的工作和生活中克服各种困难，不断实现自我的人
大学生个性简单的自我介绍经典范文 大学生的能够让在进行自我介绍的时候脱颖而出。那么关于个性简单自我介绍的范文有哪些呢?下面为大家了大学生个性简单的自我介绍经典范文，希望大家喜欢。 Hi,My name is Jay,I'm from the beautiful ancient city of Kaifeng.As you can see,I am a very casual girl,and a lot of people here,like 18-year-old,I love a lot,I love guitar,love to sing love dancing,very fond of English,I am very love to watch "Prison Break",like the actor micheal clever wit.I like making friends,and hope you will be able to and I have bee good friends,I think I will,and you get along with. From a middle class family,I was born in Hsin Ying,Tainan on October 10th,1965.My father is a civil official at Tainan City ___.My mother is a house wife good at cooking.Although I am the only child of my parents,I am by no mans a spoiled one.On the contrary,I have been expected to be a suessful man with advanced education.I study hard at school.Besides texts knowledge,journalism is
大学开学自我介绍简洁 大家好，我叫 XXX，大家可以叫我的外号 XX。 我毕业于 xx 学校(或我来自 xx-x)，正如大家一样，在见到如此多的新面孔的时候，我也 激动万分。 我的性格，我的爱好俗话说：有缘千里来相会，无缘对面不相识。 因此我觉得能跟大家在一起学习是一种缘分，希望在以后的日子里我们互相帮助共同进 步，更上一层楼。 谢谢! 文艺版大家好，我叫 xx-x，毕业于 XX 高中 。 我从小便喜好文学，对我而言，相比理科中奇妙的符号数字和多变的几何图形，我更喜爱 文学中的那一片人文气息。 无人时，我喜欢静静的看书，沉浸甚至沉醉于那一部部文学名著中。 书中那片浓郁的人文气息，沁人心脾，让我不由的深陷其中，无法自拔。 不过由于过于以书为友，反而让我忽略了现实中的人际交流，在面对陌生的环境和陌生的 人时，我会由于茫然不知所措而显得过于沉默乃至沉闷，这也算是我最大的一个缺点吧，所以 我希望在大学里多和同学们交流，努力改掉这个缺点。 在今后的四年里，我希望在收获好成绩的同时，也能收获到数份珍贵的友谊! 在此先谢谢 大家了! 随和版大家好，我是 xx-x，来到这里，我非常的兴奋，因为我又能在新的环境中， 找到新的朋友了。 我是一名学习一般，相貌一般，体格一般的内向学生。 我沉着冷静，比较和善，也比较好相处，大家可以和我多交朋友。 我喜欢看书，特别是喜欢 xx-x 的书，希望在大学能够交到一些志同道合的书友。 我和在座的同学们一样，渴望展翅高飞，渴望将来有更大的发展空间，有施展才华的更广 阔的天地。 我想，有耕耘就会有收获。 未来的四年里，由各位老师的倾情传授，我们一定会有一个更加无限美好的未来。 谢谢大家。 幽默版 女生 hello，大家好，我叫 xx-x，小女芳龄 18 岁。 我的小毛病很多，天知道我们会被缘分紧紧缠在一起，好吧，你们呢就认了吧，注定我们 是朋友了!，“高调做事，低调做人是我的做人准则，所以有时候会小矫情，还希望大家多多 关照。 小女子我性格活泼开朗(幽默的孩子的共性，呵呵)，爱好广泛，喜爱交友。 大学四年的时间说长也不长说短也不短，我们会看到彼此的颓唐或者进步，好了不说了， 总之一句话——(此处顿一下)此时不搏何时搏。 男生嘿，大家好，我是 xx-x。 青春年少的我，和大部分男孩子一样喜欢运动，最爱篮球(什么爱好不重要)。
大数据平台构思方案 （项目需求与技术方案） 一、项目背景 “十三五”期间，随着我国现代信息技术的蓬勃发展，信息化建设模式发生根本性转变,一场以云计算、大数据、物联网、移动应用等技术为核心的“新 IT”浪潮风起云涌，信息化应用进入一个“新常态”。***（某政府部门）为积极应对“互联网+”和大数据时代的机遇和挑战，适应全省经济社会发展与改革要求，大数据平台应运而生。 大数据平台整合省社会经济发展资源，打造集数据采集、数据处理、监测管理、预测预警、应急指挥、可视化平台于一体的大数据平台，以信息化提升数据化管理与服务能力，及时准确掌握社会经济发展情况，做到“用数据说话、用数据管理、用数据决策、用数据创新”，牢牢把握社会经济发展主动权和话语权。 二、建设目标 大数据平台是顺应目前信息化技术水平发展、服务政府职能改革的架构平台。它的主要目标是强化经济运行监测分析，实现企业信用社会化监督，建立规范化共建共享投资项目管理体系，推进政务数据共享和业务协同，为决策提供及时、准确、可靠的信息依据，提高政务工作的前瞻性和针对性，加大宏观调控力度，促进经济持续健康发
展。 1、制定统一信息资源管理规范，拓宽数据获取渠道，整合业务信息系统数据、企业单位数据和互联网抓取数据，构建汇聚式一体化数据库，为平台打下坚实稳固的数据基础。 2、梳理各相关系统数据资源的关联性，编制数据资源目录，建立信息资源交换管理标准体系，在业务可行性的基础上，实现数据信息共享，推进信息公开，建立跨部门跨领域经济形势分析制度。 3、在大数据分析监测基础上，为政府把握经济发展趋势、预见经济发展潜在问题、辅助经济决策提供基础支撑。 三、建设原则 大数据平台以信息资源整合为重点，以大数据应用为核心，坚持“统筹规划、分步实施，整合资源、协同共享，突出重点、注重实效，深化应用、创新驱动”的原则，全面提升信息化建设水平，促进全省经济持续健康发展。