<![CDATA[Deaf people use sign language as their primary means of communication. Images of sign language gestures are usually low quality because visual impairments like noise and low contrast prevent an automatic recognition system from working well. This research tries to enhance the quality of images with sign language gestures using two preprocessing methods, namely Gaussian Filter and Contrast Stretching. The first one eliminates noise while keeping important details in the image, and the second increases pixel intensity distribution to make hand gestures more apparent and outlined. An experiment was done on a dataset that includes 54,049 static hand gesture images taken from videos that contain certain sign languages divided into 28 classes for hijaiyah letters. A quantitative evaluation indicated substantial enhancements in processed image quality. The preprocessing method resulted in an average PSNR of 20.13 dB, SSIM equal to 0.8875, and MSE equal to 976.39 for all samples tested confirming that this combination method improves sharpness, structural integrity, and contrast when compared with original unprocessed images significantly. This study recommends using Gaussian Filter along with Contrast Stretching as a practical option for improving the quality of sign language images which can eventually help automated recognition systems that need clear visual input to correctly classify gestures.]]>
Copyrights © 2025
