A Degradation Type Adaptive and Deep CNN-Based Image Classification Model for Degraded Images

• Published in: Computer modeling in engineering & sciences Vol. 138; no. 1; pp. 459 - 472
• Main Authors: Liu, Huanhua, Wang, Wei, Liu, Hanyu, Yi, Shuheng, Yu, Yonghao, Yao, Xunwen
• Format: Journal Article
• Language: English
• Published: Henderson Tech Science Press 2024
• Subjects: Accuracy; Artificial neural networks; Classification; Differential thermal analysis; Image classification; Image degradation; Image quality; Performance evaluation

Deep Convolutional Neural Networks (CNNs) have achieved high accuracy in image classification tasks, however, most existing models are trained on high-quality images that are not subject to image degradation. In practice, images are often affected by various types of degradation which can significan...