Variable bit allocation method based on meta-heuristic algorithms for facial image compression

High spatial resolution is one of the most important factors in increasing image quality, but it increases the amount of storage memory. On the other hand, in the field of face image compression research, one of the existing challenges is maintaining the image recognition rate. Therefore, it would be very desirable to propose a method that does not reduce the minimum detection rate. This paper examines how to compress face images with high spatial resolution based on meta-heuristic algorithms. In such a way that the images are divided into the same blocks and the task of identifying the important blocks is the responsibility of meta-heuristic algorithms. The fitting function in this condition is to achieve the maximum value of image recognition accuracy. In the simulation and evaluation section, face images of CIE and FEI databases have been investigated as a selective study. The simulation results show the significant effect of the proposed methods using meta-heuristic algorithms in increasing the quality of PSNR and SSIM compared to the detection efficiency. The recognition accuracy of face images shows 1.25% more than the condition without compression for the gray wolf meta-heuristic algorithm. According to the proposed method, the larger the block division value, the better the average PSNR and SSIM. In general, depending upon the application of the problem, there is a trade-off to achieve the highest average PSNR or SSIM using genetic, whale, or gray wolf algorithms. However, the gray wolf algorithm reaches its optimal answer much faster than the genetic and whale algorithms.