A Hybrid Approach to Image Fusion Using Wavelet Transform and Weighted Arithmetic Optimization

The exponential increase in digital imagery has amplified the demand for effective tools to identify similar images and merge them into richer informative outputs. with the applications as crossing image Enhancement, digital forensic science and content management. This work proposes an Advanced Image Similarity Detection and Fusion approach that combines computer vision, wavelet transform and adaptive optimization to respond for such demands. It adopts histogram and edge-based feature extraction and cosine similarity to find the similar image pairs from a folder with robust detection by the equipping a configurable threshold (default: 0.85). It uses multi-level wavelet decomposition with the symlet 8 wavelet technique for fusion, adaptively weights coefficients from energy variance, and reconstructs images using unsharp masking for better clarity. Implemented in Python using libraries such as OpenCV, PyWavelets, and NumPy, the system handles various image formats (PNG, JPG, JPEG, BMP) in real-time, producing fused outputs with metadata. This work provides a scalable, highfidelity solution that surpasses conventional approaches, making a contribution to the area of image processing by achieving accuracy, quality, and efficiency.