Research And Applications On Image Feature Extraction

Feature extraction is always one of the most typical research topics in the image processing and computer vision. It has been taken widely use of in the areas of military, computer sciences, computer-aided diagnose, industry and human-machine-communication. Feature extraction has a wide range of applications and potential market value.Active Shape Models(ASM) is one of powerful tools for feature extraction.However, the performance of ASM is often influenced by some factors such as the initial location, illumination and so on, which will frequently lead to the local minima in optimization.In this paper, classical Active Shape Model (ASM) is studied including shape matching algorithm based on gray-level. Some factors which impact ASM searching are conducted and analyzed.The performances of our methods are illustrated by simulation experiment results. The major contributions of this paper are as follows.(1) This paper proposes a weighted Active Shape Models, in which the more robust local appearance model is constructed with the local information of each landmark fully.(2) Artificial fish-swarm algorithm is an animal's autonomous method that bases on the principle of artificial intelligent. It has some characters, such as no special requirement for object functions, being insensitive to the initial values, tolerating wide range of values of parameters, having the abilities of parallel processing and global search. This paper introduced an ASM based on artificial fish-swarm algorithm.Finally, those improved methods are evaluated on ORL face database and JAFFE face database. Through the improved method the local minima problem can be solved efficiently, extracted the detailed local information of feature points and localized the points accurately. The experimental results show that the new algorithms have some advantages in terms of searching speed, searching accuracy, avoiding local optimum.