Super-Resolution of Color Images Using CNN
Ashvini A. Tandale, Nitiraj V. Kulkarni · 2018
Convolutional neural networks (CNN) offer superior performance for Single Image Super Resolution (SISR) tasks. Super-resolution is a technique that improves low-resolution image quality and converts it into high-resolution images to provide better viewing. As the network grows, the features of the previous levels are prevented or not used in subsequent levels. In SISR, the previous layers are mainly composed of local characteristics that are essential for the activity. The proposed super resolution CNN approach is developed for x3 and x4 scaling factor. The proposed approach achieves the good results.