Enhancing image compression efficiency using DCT and computational intelligence-based optimization

Traditional JPEG compression comes with a significant limitation. Its rigid quantization methods and fixed entropy coding often make it challenging to achieve the right balance between image quality and compression efficiency. To tackle this issue, our paper introduces an innovative image compression system that enhances the standard JPEG approach by incorporating computational intelligence in two crucial stages. We've created a flexible quantization process that uses a genetic algorithm (GA) as its guiding framework. This approach enables us to generate a customized quantization matrix for every single image, leading to enhanced adaptability and better performance when it comes to rate-distortion. Secondly, we revamped the coding process by implementing a dynamic cross-entropy coding method. This mechanism intelligently toggles between Huffman and arithmetic coding depending on which approach offers better coding efficiency during testing. This hybrid approach significantly reduces the mean squared error (MSE) and improves the peak signal-to-noise ratio (PSNR) while maintaining a good compression ratio in our experiments with standard grayscale images. These results demonstrate that our novel method improves image quality and reduces storage requirements by adding flexibility to the JPEG format.