This project helps to improve the compression techniques for images using highly efficient methods such as AAM (Adjacent Attention Modules) and RM (Refinement Blocks). The proposed approach of the project helps in striking a balance between maintaining the image quality and compression efficiency. Older compression techniques are used to struggle for preserving the important details without making compromises in smaller file sizes. This project enhances the image quality by efficiently using the AAM for the process for procuring the long-range focused dependencies present within the images, without making any compromises in preserving the important and spatial information during the process of compression. We will be additionally using the refinement blocks for refining the images that are compressed using AAM for enhancing the important and critical features of the respective images. Through repeated experimentation and rigorous evaluation using high-standard efficient datasets, the effectiveness and efficiency of this project was demonstrated. These blocks are tailored to capture local content, optimize rate-distortion, and refine compressed images post-compression, thereby enhancing overall compression performance.