In the field of image enhancement, using deep learning methods to enhance low-light images is currently mainstream. However, the methods often have complex network structures of a large number of parameters, and their training often uses paired data-sets, which are difficult to obtain in actual practice. To solve these problems, this paper proposes a simple generative adversarial network and Retinex model, dubbed RetinexGAN, that is completely trained using unpaired data-sets. It contains a decomposition network and two discriminator networks. To reduce the parameters of the network, only two convolution layers are used in the decomposition network. We show more challenging testing data where some parts of the image are underexposed and others are normal light. Both quantitative and visual results show that RetinexGAN is largely superior to state-of-the-art methods.

How to generate the path planning of mobile robots quickly is a problem in the field of robotics. The Q-learning(QL) algorithm has recently become increasingly used in the field of mobile robot path planning. However, its selection policy is blind in most cases in the early search process, which slows down the convergence of optimal solutions, especially in a complex environment. Therefore, in this paper, we propose a continuous local search Q-Learning (CLSQL) algorithm to solve these problems and ensure the quality of the planned path. First, the global environment is gradually divided into independent local environments. Then, the intermediate points are searched in each local environment with prior knowledge. After that, the search between each intermediate point is realized to reach the destination point. At last, by comparing other RL-based algorithms, the proposed method improves the convergence speed and computation time while ensuring the optimal path.

Image super-resolution reconstruction is one of the methods to improve resolution by learning the inherent features and attributes of images. However, the existing super-resolution models have some problems, such as missing details, distorted natural texture, blurred details and too smooth after image reconstruction. To solve the above problems, this paper proposes a Multi-scale Dual-Attention based Residual Dense Generative Adversarial Network (MARDGAN), which uses multi-branch paths to extract image features and obtain multi-scale feature information. This paper also designs the channel and spatial attention block (CSAB), which is combined with the enhanced residual dense block (ERDB) to extract multi-level depth feature information and enhance feature reuse. In addition, the multi-scale feature information extracted under the three-branch path is fused with global features, and sub-pixel convolution is used to restore the high-resolution image. The experimental results show that the objective evaluation index of MARDGAN on multiple benchmark datasets is higher than other methods, and the subjective visual effect is better. This model can effectively use the original image information to restore the super-resolution image with clearer details and stronger authenticity.