Enhancement of Degraded Image Based on Neural Network
Defa Hu, Zhuang Wen Wu · 2015
Inevitably, image degradation is caused in the imaging, reproduction, scanning, transmission and display. This paper conducts systematic analysis and research on the clearness processing of degraded image based on BP neural network. The proposed scheme increases the output ranges of the hidden layer and the output layer, sets the variable step size, accelerates the learning speed of the neural network and avoids the excessive correction of the weights and speeds up the convergence rate of the neural network. It can preserve more profound image information, greatly enhance the contrast of the degraded image, effectively strengthen the overall quality of the degraded image and obtain more ideal image clearness effect while processing the degraded image acquired under bad conditions.