MULTIFRAME IMAGE RESTORATION - ENHANCING IMAGE QUALITY THROUGH ADVANCED RECONSTRUCTION TECHNIQUES

The degradation of image quality due to noise, blur, and low contrast remains a significant challenge in various imaging applications, particularly in medical diagnostics, remote sensing, and surveillance. Effective restoration of such images is essential to enhance visual clarity and extract meaningful information. Conventional techniques often struggle to balance noise reduction and detail preservation. To address these limitations, this study proposes an advanced multiframe image restoration approach combining Contrast Limited Adaptive Histogram Equalization (CLAHE) and Deep Belief Networks (DBN). CLAHE is employed to enhance contrast adaptively, improving visibility in regions with varying luminance. Subsequently, DBN, a deep learning model, is applied to refine the reconstruction process by leveraging its feature extraction and noise suppression capabilities. This combination ensures that the restored images retain fine details while effectively mitigating noise and distortions. Experimental evaluation was conducted on a dataset of 500 degraded images, including medical scans and natural scenes. The proposed method achieved a Peak Signal-to-Noise Ratio (PSNR) of 36.2 dB, a Structural Similarity Index (SSIM) of 0.92, and a contrast improvement rate of 48%, surpassing traditional methods like Bilateral Filtering and Wavelet Transform. Processing time per image was maintained at an efficient 1.8 seconds, ensuring practicality for real-time applications. This novel integration of CLAHE and DBN shows significant advancements in multiframe image restoration, making it a valuable tool for applications requiring enhanced image quality. The approach combines the strengths of contrast enhancement and deep learning- based reconstruction, paving the way for improved image analysis and decision-making in critical domains.