Current Issue : January - March Volume : 2021 Issue Number : 1 Articles : 5 Articles
Online education has developed rapidly due to its irreplaceable convenience. Under the severe circumstances caused by COVID-\n19 recently, many schools around the world have delayed opening and adopted online education as one of the main teaching\nmethods. However, the efficiency of online classes has long been questioned. Compared with traditional face-to-face classes, there\nis a lack of direct, timely, and effective communication and feedback between teachers and students in the online courses. Previous\nstudies have shown that there is a close and stable relationship between a personâ??s facial expressions and emotions generally. From\nthe perspective of computer simulation, a framework combining a face expression recognition (FER) algorithm with online\ncourses platforms is proposed in this work. The cameras in the devices are used to collect studentsâ?? face images, and the facial\nexpressions are analyzed and classified into 8 kinds of emotions by the FER algorithm. An online course containing 27 students\nconducted on Tencent Meeting is used to test the proposed method, and the result proved that this method performs robustly in\ndifferent environments. This framework can also be applied to other similar scenarios such as online meetings....
This article reports the results of the study related to emotion recognition by using eye-tracking. Emotions were evoked by\npresenting a dynamic movie material in the form of 21 video fragments. Eye-tracking signals recorded from 30 participants were\nused to calculate 18 features associated with eye movements (fixations and saccades) and pupil diameter. To ensure that the\nfeatures were related to emotions, we investigated the influence of luminance and the dynamics of the presented movies. Three\nclasses of emotions were considered: high arousal and low valence, low arousal and moderate valence, and high arousal and high\nvalence. A maximum of 80% classification accuracy was obtained using the support vector machine (SVM) classifier and leaveone-\nsubject-out validation method....
Computer vision is one of the hottest research fields in deep learning. The emergence of generative adversarial networks (GANs)\nprovides a new method and model for computer vision. The idea of GANs using the game training method is superior to\ntraditional machine learning algorithms in terms of feature learning and image generation. GANs are widely used not only in\nimage generation and style transfer but also in the text, voice, video processing, and other fields. However, there are still some\nproblems with GANs, such as model collapse and uncontrollable training. This paper deeply reviews the theoretical basis of GANs\nand surveys some recently developed GAN models, in comparison with traditional GAN models. The applications of GANs in\ncomputer vision include data enhancement, domain transfer, high-quality sample generation, and image restoration. The latest\nresearch progress of GANs in artificial intelligence (AI) based security attack and defense is introduced. Thefuture development of\nGANs in computer vision is also discussed at the end of the paper with possible applications of AI in computer vision....
Object tracking based on low-rank sparse learning usually makes the drift phenomenon occur when the target faces severe\nocclusion and fast motion. In this paper, we propose a novel tracking algorithm via reverse low-rank sparse learning and\nfractional-order variation regularization. Firstly, we utilize convex low-rank constraint to force the appearance similarity of the\ncandidate particles, so as to prune the irrelevant particles. Secondly, fractional-order variation is introduced to constrain the sparse\ncoefficient difference in the bounded variation space, which allows the difference between consecutive frames to exist, so as to\nadapt object fast motion. Meanwhile, fractional-order regularization can restrain severe occlusion by considering more adjacent\nframes information. Thirdly, we employ an inverse sparse representation method to model the relationship between target\ncandidates and target template, which can reduce the computation complexity for online tracking. Finally, an online updating\nscheme based on alternating iteration is proposed for tracking computation. Experiments on benchmark sequences show that our\nalgorithm outperforms several state-of-the-art methods, especially exhibiting better adaptability for fast motion and\nsevere occlusion....
Semantic segmentation with convolutional neural networks under a complex background using the encoder-decoder network\nincreases the overall performance of online machine vision detection and identification. To maximize the accuracy of semantic\nsegmentation under a complex background, it is necessary to consider the semantic response values of objects and components\nand their mutually exclusive relationship. In this study, we attempt to improve the low accuracy of component segmentation. The\nbasic network of the encoder is selected for the semantic segmentation, and the UPerNet is modified based on the component\nanalysis module. The experimental results show that the accuracy of the proposed method improves from 48.89% to 55.62% and\nthe segmentation time decreases from 721 to 496ms. The method also shows good performance in vision-based detection of 2019\nChinese Yuan features....
Loading....