Dissertation > Excellent graduate degree dissertation topics show

Research on Shot Retrieval Based on Fuzzy Evolutionary aiNet and Probabilistic Distance

Author: LiXianHui
Tutor: ZhanYongZhao
School: Jiangsu University
Course: Applied Computer Technology
Keywords: Shot Retrieval Fuzzy evolutionary immune network Probability distance Spatial and temporal characteristics Shot similarity
CLC: TP391.41
Type: Master's thesis
Year: 2009
Downloads: 36
Quote: 0
Read: Download Dissertation

Abstract


With the rapid development of multimedia technology, multimedia data increase exponentially. Therefore, how fast and efficient from the vast amounts of video retrieved the required video becomes very important. Content-based video retrieval via video structure analysis, the video is divided into key frames, lenses, scene final lens units, according to user-submitted video instance, in the video database Find similar video clips, and based on the similarity degree the level given in the search results. Meet the above requirements to a certain extent. This article first summarizes the research background and analysis at home and abroad have been proposed mainstream content-based video retrieval method. Comprehensive analysis based video retrieval method based on this paper, the key frame extraction algorithm based on fuzzy evolutionary immune network, the lens spatial and temporal feature extraction methods, and the similarity measure based on the the probability distance the lens, as follows: (1) proposes a keyframe extraction algorithm based on fuzzy evolutionary immune network. By fuzzy evolutionary immune network theory is introduced into the proposed algorithm based on fuzzy evolutionary immune network key frame extraction. The algorithm to each frame of a lens as antigen Ag the fuzzy evolutionary immune network node (extracted keyframes) as antibody Ab and continuously optimized by on fuzzy evolutionary immune network update, eventually extracted the most key able to represent the whole of the lens frame or a combination of key frames. (2) study the lens, the empty feature extraction methods. Time characteristics define the lens with a time weighted color histogram to represent each lens corresponds to a lens weighted color histogram; spatial characteristics, on behalf of the significance of the five parameters by calculating GLCM entropy, contrast, energy, correlation and inverse gap to indicate the characteristics of an image texture; respectively calculate the standard deviation of the area ratio, centroid, x direction and y direction, x direction by the block diagram of the color of the frame image for each block diagram constructed in and regional aspect ratio in the y direction to the spatial configuration information statistical histogram represents the structural characteristics of the image space. (3) propose a similarity measure based on probability distance lens. Underlying spatial feature vector of the similarity in terms of the spatial information, the introduction of the theory of probability distance, first extracted with a lens to be regarded as a combination, then the spatial feature vector combination is mapped to a high-dimensional space by a nonlinear mapping Gaussian distribution modeling, and finally through the calculation of the probability distance between two spatial characteristics of the high-dimensional space vector Gaussian distribution to measure the similarity of two lens space; similar time information, each lens corresponds to a lens weighted color histogram using histogram intersection The method for solving; ultimately, the overall similarity of the two lenses can be adopted to solve similarity weighted sum of the similarity and the time information of the lens space. (4) research and development of the lens retrieval prototype system framework. The modular design concept, design shot retrieval prototype system to verify the validity of the above method and experimental comparison.

Related Dissertations

  1. Temporal and Spatial Distribution of the Interannual Variability of the Global Precipitation in the Tropical Region,P426.613
  2. Song Dynasty money monitoring research,K244
  3. The Great Wall during the Spring and Autumn and the Warring States Period in Spatial-temporal Perspective,K878
  4. Study on the Optimal Allocation of Human Resources in Liaoning Coastal Economic Zone,F249.27
  5. Study on Distribution of Polycyclic Musks in Water of Songhua River,X703
  6. Study on the Temporal-Spatial Features of Intensified Cultivated Land Use in Jiangsu Province,F301.21
  7. Research on Remote Sensing Monitoring of Urban Spatial Expansion in Inner Mongolia,P237
  8. Research on the Key Technique of Simple Action Recognition in Video Sequence,TP391.41
  9. The Effect of Physiological, Psychological and Technical on Pole Vaulters under Psychological Regulating,G804.2
  10. Research on Simulation of Ultra-short Laser Pulse Spatio-temporal Characteristics,TN249
  11. Spatio-temporal Patterns and Comprehensive Clustering Analysis of Climatic Change Risk Sources in Qinghai-Tibet Plateau,P467
  12. Analysis of Lightning Characteristics Based on SAFIR Data in Beijing and Its Circumjacent Regions,P427.32
  13. A Study on Winter Snow and Rain in Northern China,P468.0
  14. Identification and Distribution Characteristics of Sea and Land Breeze in the Western Bohai,P412.16
  15. Improved singular value decomposition method and its application,P434
  16. Fujian bamboo forest ecosystem carbon sinks Seasonal Variation,S718.5
  17. The Study of Vegetable Price Fluctuation and Its Relationship with the Climate Based on Wavelet Analysis,F323.7;F224
  18. Tempeoral and Spatial Distribution of Haze in China and Its Association with Climate Change,P427.21
  19. Analysis on Temporal and Spatial Distribution of Ocean Wind Around the Typical Routes of the Ocean Ship,U692
  20. A Study of Urban Air Quality Prediction in Nanning City Base on Combination of the Genetic Algorithm and Artificial Neural Networks,X823

CLC: > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Computer applications > Information processing (information processing) > Pattern Recognition and devices > Image recognition device
© 2012 www.DissertationTopic.Net  Mobile