<p>Images are vital for understanding and processing information. However, limitations such as storage space make image compression essential. Image compression involves removing redundant information while retaining critical data to reconstruct the image effectively. In this study, we introduce a novel hybrid method for image compression that combines wavelet transforms with deep learning techniques. Our approach utilizes an encoder-decoder architecture to compress and reconstruct images efficiently. In the encoder phase, key features are extracted, and in the decoder phase, these features are used to accurately reconstruct the image. Unlike traditional methods that use max pooling, which can lead to the loss of important information, we employ wavelet transforms to resize feature maps, leading to the preservation of crucial details, ensuring the retention of spatial features, and improving the performance. To demonstrate the superiority of our approach, we compare it against state-of-the-art methods across various datasets. The experimental results clearly show that our method outperforms other methods, highlighting its effectiveness.</p>