Anew Improved Tactic to Extract Facial Expression Based on Genetic Algorithm and WVDF

Md Alamgir Hossain
2012 International Journal of Advanced Information Technology  
The study and analysis of facial expression has paying attention tremendously as human being is very conscious in the domain of computer vision because it takes a principal role in the area of man-machine communication. Almost all types of face expressions are basically spawned by taking a new hue of facial muscles, which grades in temporally malformed facial features such as forehead, eye-lids, eye-brows, nose, lips, and skin colour/texture, etc. During the evaluation procedure of any facial
more » ... ure of any facial expression emphasise should have to be given on the following parameters :( a) the parts of the face that will play prime role in facial action,(b) the intensity of facial actions, and (c) the stochastic parameters of facial actions. In our paper we have proposed a real-time, precise-exact, and robust facial expression recognition methodology depending on genetic algorithm along with Weighted Vector Directional Filters (WVDF) as it is proved that it is the most powerful unbiased optimization techniques for sampling a huge solution space. The GA can be optimized by using WVDF filter's coefficients. By using the Stochastic Selection of parents and the generation of optimal offspring, the optimal WVDF filter can be obtained. We have been capable to construct a leaf-matrix to our required intended expression. Ultimately, we have evaluated our proposed methodology by using Weighted Vector Directional Filters (WVDF) along with genetic algorithm with the shutter function of Canon Camera. We have accomplished our job by achieving better performance than Sony on slight function.
doi:10.5121/ijait.2012.2504 fatcat:sm3lzyjdize7jeup3uim2tyh7a