The appearance of generative adversarial networks (GAN) provides a new approach and framework for computer vision. Compared with traditional machine learning algorithms, GAN works via adversarial training concept and is more powerful in both feature learning and representation. GAN also exhibits some problems, such as non-convergence, model collapse, and uncontrollability due to high degree of freedom. How to improve the theory of GAN and apply it to computer-vision-related tasks have now attracted much research efforts. In this paper, recently proposed GAN models and their applications in computer vision are systematically reviewed. In particular, we firstly survey the history and development of generative algorithms, the mechanism of GAN, its fundamental network structures, and theoretical analysis of the original GAN. Classical GAN algorithms are then compared comprehensively in terms of the mechanism, visual results of generated samples, and Frechet Inception Distance. These networks are further evaluated from network construction, performance, and applicability aspects by extensive experiments conducted over public datasets. After that, several typical applications of GAN in computer vision, including high-quality samples generation, style transfer, and image translation, are examined. Finally, some existing problems of GAN are summarized and discussed and potential future research topics are forecasted.

Due to the complexity of myocardial infarction (MI) waveform, most traditional automatic diagnosis models rarely detect it, while those able to detect MI often require high computing and storage capacity, rendering them unsuitable for portable devices. Therefore, in order for convenient real-time MI detection, it is essential to design lightweight models suitable for resource-limited portable devices. This paper proposes a novel multi-channel lightweight model (ML-Net), that provides a new solution for portable detection devices with limited resources. In ML-Net, each electrocardiogram (ECG) lead is assigned an independent channel, ensuring data independence and preserve the ECG characteristics of different angles represented by different leads. Moreover, convolution kernels of heterogeneous sizes are utilized to achieve accurate classification with only a small amount of lead data. Extensive experiments over actual ECG data from the PTB diagnostic database are conducted to evaluate ML-Net. The results show that ML-Net outperforms comparable schemes in diagnosing MI, and it requires lower computational cost and less memory, so that portable devices can be more widely used in the field of Internet of Medical Things(IoMT).

By running applications and services closer to the user, edge processing provides many advantages, such as short response time and reduced network traffic. Deep-learning based algorithms provide significantly better performances than traditional algorithms in many fields but demand more resources, such as higher computational power and more memory. Hence, designing deep learning algorithms that are more suitable for resource-constrained mobile devices is vital. In this paper, we build a lightweight neural network, termed LiteNet which uses a deep learning algorithm design to diagnose arrhythmias, as an example to show how we design deep learning schemes for resource-constrained mobile devices. Compare to other deep learning models with an equivalent accuracy, LiteNet has several advantages. It requires less memory, incurs lower computational cost, and is more feasible for deployment on resource-constrained mobile devices. It can be trained faster than other neural network algorithms and requires less communication across different processing units during distributed training. It uses filters of heterogeneous size in a convolutional layer, which contributes to the generation of various feature maps. The algorithm was tested using the MIT-BIH electrocardiogram (ECG) arrhythmia database; the results showed that LiteNet outperforms comparable schemes in diagnosing arrhythmias, and in its feasibility for use at the mobile devices.

Body movement is one of the most important indicators of sleep quality for elderly people living alone. Body movement is crucial for sleep staging and can be combined with other indicators such as breathing and heart rate to monitor sleep quality. Nevertheless, traditional sleep monitoring methods are inconvenient and may invade users’ privacy. To solve these problems, we propose contactless Body Movement Recognition (CBMR) method via WiFi signals. Firstly, CBMR uses commercial off-the-shelf WiFi devices to collect Channel State Information (CSI) data of body movement and segment the CSI data by sliding window. Then, the context information of the segmented CSI data is learned by a Bi-directional Recurrent Neural Network (Bi-RNN). Bi-RNN can fuse the forward and backward propogation information at some point, and input it into a deeper Independently Recurrent Neural Network (IndRNN) with residual mechanism to extract the deeper features and capture the time dependencies of CSI data. Finally, the type of body movement can be recognized and classified by the softmax function. CBMR can effectively reduce data preprocessing and the delay caused by manually extracting features. The results of an experiment conducted on a complex body movement dataset show that our method gives desirable performance and achieves average accuracy of greater than 93.5%, which implies a prospect application of CBMR.

Free viewpoint video (FVV) offers compelling interactive experience by allowing users to switch to any viewing angle at any time. An FVV is composed of a large number of camera-captured anchor views, with virtual views (not captured by any camera) rendered from their nearby anchors using techniques such as depth-image-based rendering (DIBR). We consider a group of wireless users who may interact with an FVV by independently switching views. We study a novel live FVV streaming network where each user pulls a subset of anchors from the server via a primary channel. To enhance anchor availability at each user, a user generates network-coded (NC) packets using some of its anchors and broadcasts them to its direct neighbors via a secondary channel. Given limited primary and secondary channel bandwidths at the devices, we seek to maximize the received video quality (i.e., minimize distortion) by jointly optimizing the set of anchors each device pulls and the anchor combination to generate NC packets. To our best knowledge, this is among the first body of work addressing such joint optimization problem for wireless live FVV streaming with NC-based collaboration. We first formulate the problem and show that it is NP-hard. We then propose a scalable and effective algorithm called PAFV (Peer-Assisted Freeview Video). In PAFV, each node collaboratively and distributedly decides on the anchors to pull and NC packets to share so as to minimize video distortion in its neighborhood. Extensive simulation studies show that PAFV outperforms other algorithms, achieving substantially lower video distortion (often by more than 20-50%) with significantly less redundancy (by as much as 70%). Our Android-based video experiment further confirms the effectiveness of PAFV over comparison schemes.