Enhancing Diagnostic Accuracy Denoising Medical Imaging
Sallauddin Mohmmad, Nagurla Mahender, Bharath Aaleti, Adupa Shashank, R Niveditha, Mohammed Zeeshan Azeez Ur Rahman · 2024
Medical image processing is faced with the challenge of accurate diagnosis founded on image noise, blurriness, and lack of proper preprocessing. In this paper, our model introduced various denoising techniques in medical image processing, from customary methods like Non-Local Means and Wavelet-Based Denoising to modern approaches such as autoencoders and the resultant images classified using the U-Net architecture. Autoencoders, akin to smart filters, exhibit assurance in efficiently annihilating unwanted artifacts while holding essential image features. The model's primary goal is to leverage autoencoders to enhance the clarity and Precision of medical image data and address challenges originating from image copying and quality loss. By utilizing U-Net neural networks, this analysis achieved an accuracy of 97.15%, Precision (97.87%), and recall (96.67%).