RIFD-LZW: A Hybrid Approach for Lossy Image Compression Using Intensity Rounding, Division, and Lempel-Ziv-Welch Encoding

This research presents RIFD-LZW, a new hybrid lossy image compression algorithm designed for both color and grayscale images across varying resolutions. The method integrates the Rounding the Intensity and Dividing (RIFD) technique with Lempel-Ziv-Welch (LZW) encoding to enhance compression efficiency while preserving high image quality. The RIFD stage reduces data redundancy through intensity quantization and scaling, while LZW applies efficient lossless dictionary-based encoding to the transformed data. Comprehensive experiments were conducted on four benchmark datasets EPFL, Kodak, Waterloo, and HQ-50K to evaluate the performance of the proposed method. The results demonstrate that RIFD-LZW consistently outperforms traditional RIFD, LZW, and standard compression algorithms including JPEG2000, JPEG-LS, and RIFD-Huffman. On average, RIFD-LZW achieved a compression efficiency of 7,51 Bits Per Pixel (BPP) for color datasets, representing a 49,93% improvement over RIFD and 62,49% over LZW. For grayscale images, RIFD-LZW attained an average BPP of 1,92, significantly outperforming RIFD (5,00) and LZW (4,74), with an improvement exceeding 59%. The RIFD-LZW algorithm delivers high visual quality despite being lossy, achieving average PSNR values 38,36 dB with minimal visible distortion. It effectively reduces file sizes while preserving acceptable image quality, making it well-suited for applications that require efficient compression with good visual retention.