A Perceptually-Based Deep Learning Approach for Energy-Preserving Image Enhancement

Display-enabled devices are an integral part of our daily lives. A good consumer experience depends on prolonged usage without interruptions, making energy consumption a critical factor in designing such devices. In this paper, we propose a convolutional neural network that enhances perceptual quality while keeping energy consumption fixed, offering a solution to extend device usage time. Our approach utilizes a balanced combination of loss functions for energy consumption and perceptual quality for the training phase, enabling our model to generate energy-efficient images with enhanced quality. Subjective and objective tests demonstrate the superiority of our method in improving the quality of dimmed images compared to the state-of-the-art methods. The code and results are available at: https://github.com/AlirezaRafieiSardouei/PerceptDimmed