Exploring the Impact of Image Quality on Convolutional Neural Networks: A Study on Noise, Blur, and Contrast

Convolutional neural networks (CNNs) have recently revolutionized several computer vision tasks, including object detection and image categorization. The secret to their astounding success is their capacity to automatically recognize and extract features from intricate data representations. However, the caliber of the input images significantly affects how well CNNs work. The critical importance of image quality in the context of CNNs is explored in depth in this study, emphasizing how blur, noise, and contrast affect the overall performance of CNNs. The classification performance of degraded images is also demonstrated by this study to be influenced by the picture quality of training data for a classification network.