Dissertation > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Computer applications > Information processing (information processing) > Pattern Recognition and devices > Image recognition device

Multiple Description Coding of 3D Meshes

Author GaoMingHui
Tutor NiLin
School University of Science and Technology of China
Course Signal and Information Processing
Keywords 3Dmesh MDC Layered Coding vertex spanning tree wavelet SPIHT
CLC TP391.41
Type Master's thesis
Year 2010
Downloads 36
Quotes 0
Download Dissertation

3D graphics data are more and more widely used in various applications, including engineering design, medical diagnosing and video gaming. In this paper we use the triangle mesh to represent the 3D mesh model, and concentrate on the compression of connectivity and geometry data.As the number and complexity of existing 3D meshes increase explosively, a huge amount of storage space and transmission bandwidth is needed. Unfortunately, there are not only a lot of error, packets losing in the internet and the intranet, but the users who need to access the mesh data are also equipped with personal computers and standard telephone line connections. So the vertex spanning tree and wavelet translation are used to reduce the connectivity information and the correlation of the 3D mesh, in the same time the MDC coding method is used to solve the problem of error and packets losing. The MDC encoder generates two (or more) descriptions having equal bit-rates and importance. The descriptions are packaged independently and sent over either same or separate channels. As long as the two (or more) descriptions are lost simultaneously, the MDC decoder can make a reconstruction. If only one of the descriptions is received, the MDC decoder decodes the received description using the side decoder and reconstructs the data with low but acceptable quality.Fully analyzed as above, We propose the LMDC(Layered Multiple Description Coding) witch combined the advantages of both layered and multiple description coding. First, the mesh was decomposed to derive a base mesh and the reconstruction information. Adopting the idea of layered coding, the base mesh was encoded as the base layer and the reconstruction information as the enhance layer. The MDC method based on vertex spanning tree was used to ensure that the clients can receive an acceptable base mesh. By using LMDC, it is more appropriate to the current network of real-time mesh transmission and the pursuit of higher quality. And we also propose the WW_MDC (weighted wavelet Multiple Description Coding) method based on the dependency of wavelet generated in the remeshing process. The wavelet was used to extract a base mesh and obtain an array of wavelet coefficients. Then the wavelet coefficients were scaled up before using the embedded character of SPIHT to derive multiple descriptions of the mesh, which the number of descriptions and redundant can vary according to the change of channel bandwidth and the PLR (packet loss rate). The experimental results show that the proposed methods has better ability to adapt the complex network environment and error resilient, and also lower reconstruction error.

Related Dissertations
More Dissertations