Zetao Hu, Haitao Li, Junhu Zhang, Dechun Zhang, Meiling Su
{"title":"Fish Target Detection Method Based on EPSA-CenterNet2","authors":"Zetao Hu, Haitao Li, Junhu Zhang, Dechun Zhang, Meiling Su","doi":"10.1145/3561613.3561620","DOIUrl":"https://doi.org/10.1145/3561613.3561620","url":null,"abstract":"At present, small target detection and target detection under complex backgrounds are still a major difficulty in the field of image target detection. However, fish image detection scenes often contain complex backgrounds such as water grass and reef, and fish form is small. In order to overcome the problem which is low accuracy of detection of small fish targets in complex backgrounds, In this paper, a Center Point Network 2 with Efficient Pyramid Split Attention (EPSA-CenterNet2) was proposed. The Network incorporated an Efficient Pyramid Split Attention Network (EPSANet) into CenterNet2 to improve small target detection accuracy in complex environments. In this paper, 149 images of the oplegnathus punctatus were used as a dataset to train EPSA-CenterNet2 and four other mainstream target detection networks. The experimental results showed that EPSA-CenterNet2 was superior to CenterNet2, YOLOv3, YOLOv5 and SSD in the average accuracy including AP and AP50, and the number of missed targets in small target images was less. Therefore, EPSA-Centernet2 can detect fish image targets in complex backgrounds more accurately.","PeriodicalId":348024,"journal":{"name":"Proceedings of the 5th International Conference on Control and Computer Vision","volume":"40 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"120938468","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Investigating YOLOv5 for Search and Rescue Operations Involving UAVs: Investigating YOLO5","authors":"Namat Bachir, Q. Memon","doi":"10.1145/3561613.3561644","DOIUrl":"https://doi.org/10.1145/3561613.3561644","url":null,"abstract":"Mountain recreation has become more popular, with mountaineering, rock climbing, skiing, mountain biking, hiking, and mushroom picking among the most popular sports including desert safari. Despite this tendency, there is currently limited research available explaining the rise in search and rescue as well as the injuries and illnesses that entail aid in tourist-friendly areas. Deep learning has been termed as potentially effective tool for SAR applications. Even if the individual is partially veiled, a trained deep learning system can recognize them from a variety of perspectives. Existing state-of-the-art detectors such as Faster R-CNN, YOLOv4, RetinaNet, and Cascade R-CNN have been investigated in literature on various datasets to simulate rescue scenes with acceptable results. In this research, the YOLOv5L detector is investigated for further investigation on Search and rescue dataset because of its great speed and accuracy, as well as claimed small number of false detections. The results illustrate the highest mean average accuracy and is compared with other detectors.","PeriodicalId":348024,"journal":{"name":"Proceedings of the 5th International Conference on Control and Computer Vision","volume":"39 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121208017","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Unsupervised Cross-domain Person Re-Identification based on Asymmetrical Pyramid Non-local Block","authors":"Xiangyu Li, Yuhang Zheng, Shangmin Zhou","doi":"10.1145/3561613.3561636","DOIUrl":"https://doi.org/10.1145/3561613.3561636","url":null,"abstract":"The purpose of unsupervised cross-domain (UCD) person re-identification (re-ID) is to adapt the well pre-trained model on the labeled source domain to the unlabeled target domain, which tackles a more realistic problem. However, the network in the existing model cannot fully extract the features of pedestrians, so the results after clustering are not satisfactory. To address this problem, a feature extraction network model with a self-attention mechanism is proposed in this paper in order to improve the feature expression ability. We try to design and optimize the attention mechanism-based feature extraction network and similarity loss function for unsupervised person re-ID to improve the recognition accuracy. On the basis of the baseline network (such as ResNet-50), the self-attention mechanism-asymmetrical pyramid non-local block (APNB) is added to help the network learn richer global feature representation. Besides, the similarity loss function using the Euclidean distance is designed, which shows better performance than the cosine distance. Experimental results show that the proposed method has competitive performance on two public datasets Markket-1501 and DukeMTMC-Re-ID.","PeriodicalId":348024,"journal":{"name":"Proceedings of the 5th International Conference on Control and Computer Vision","volume":"27 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117027743","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Three Dimensional Route Planning of UAVs based on Velocity Potential Field Method","authors":"Qihan Fan, Shiqiang Hu","doi":"10.1145/3561613.3561643","DOIUrl":"https://doi.org/10.1145/3561613.3561643","url":null,"abstract":"With the continuous expansion of the application field of UAV, high-speed 3D(three-dimensional) obstacle avoidance is one of the important problems that must be considered. In this paper, a new artificial potential field(APF) function is specially designed for dynamic path planning for high-speed obstacles whose speed is greater than the UAV. The theory of velocity potential field method is proposed to change the direction of resultant force, that is, add a repulsion force in the repulsion field that is perpendicular to the speed of obstacles and the speed of UAV itself. The improved velocity potential field method is simulated and compared with the original potential field method. The results show that the velocity potential field method can avoid high-speed obstacles with different speeds faster and better, and only the velocity potential field method can make the UAV avoid obstacles whose speed is greater than itself. It proves the advantage of the velocity potential field method in avoiding high-speed obstacles.","PeriodicalId":348024,"journal":{"name":"Proceedings of the 5th International Conference on Control and Computer Vision","volume":"63 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129115327","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Image-Based Storytelling Using Deep Learning","authors":"Yulin Zhu, Wei Yan","doi":"10.1145/3561613.3561641","DOIUrl":"https://doi.org/10.1145/3561613.3561641","url":null,"abstract":"In order to describe a journey, a story could be automatically generated from a group of digital photographs. Most of the existing methods focus on descriptions of specific content of a single image, such as image captioning, which lack of correlation between the images and the spatiotemporal relationships. To this end, in this paper, our goal is to propose a novel storytelling architecture based on computer vision. It makes use of visual object detection from digital images. Combining the changes in spatiotemporal domain and filling in the predetermined template, we automatically generate a text-based travel diary. In this project, compared with conventional image captioning, our aims are to effectively connect correlation between digital images and background information. The contributions of this paper are: (1) Innovative use of preset templates to generate travel diaries from photographs, associating content and context of the images as an event, (3) augmenting the images to expand the dataset, (4) shortening training time of deep learning models.","PeriodicalId":348024,"journal":{"name":"Proceedings of the 5th International Conference on Control and Computer Vision","volume":"14 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128577013","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Research on compliant plugging strategy for electric vehicle charging port based on force feedback","authors":"Wenhui He, Ruifeng Li","doi":"10.1145/3561613.3561648","DOIUrl":"https://doi.org/10.1145/3561613.3561648","url":null,"abstract":"With the continuous progress of electric vehicle intelligent driving technology and high-power fast charging technology, the automatic and unmanned electric vehicle charging modes will become the general trend. The plugging problem of the charging interface involved in the electric vehicle automatic charging technology is similar to the peg-in-hole assembly problem, but the electric vehicle charging plug and port is more complex than the structure of the general shaft and hole, which causes certain difficulties in the plugging process. Aiming at the problem that there is always a pose deviation between the charging plug and the port due to factors such as visual positioning error and environmental noise, this paper designs a compliant plugging strategy based on admittance control and a charging port search algorithm.","PeriodicalId":348024,"journal":{"name":"Proceedings of the 5th International Conference on Control and Computer Vision","volume":"11 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131206147","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Research on Classroom Interaction Behavior Analysis Algorithm based on Audio and Video","authors":"Zhiwei Zheng, Yuting Huang","doi":"10.1145/3561613.3561633","DOIUrl":"https://doi.org/10.1145/3561613.3561633","url":null,"abstract":"Classroom interaction behavior research is an important part of classroom teaching quality evaluation, which can effectively improve teaching quality. Traditional classroom interaction behavior research is mainly carried out in the form of expert lectures and student questionnaires. This method can neither make the best use of the large amount of data generated in the classroom scene, nor can it provide an objective and detailed evaluation of the teaching quality. However, in the context of educational informatization, using information technology to observe and analyze classroom interaction can make full use of teaching data and provide timely and objective feedback on the teaching situation. This paper focuses on the analysis of classroom interaction behavior in colleges and universities. In order to make full use of classroom audio and video data, a framework for classroom interaction behavior analysis based on audio and video is constructed. It divides classroom interaction behaviors into verbal and non-verbal categories, and uses deep learning technology to realize automated classroom interaction analysis. The main work and innovations are as follows: (1) Combined with the theoretical basis of traditional classroom interaction analysis and the requirements of efficient classrooms for classroom quality evaluation, this paper constructs an audio-video-based classroom interaction behavior analysis framework. (2) The speaker segmentation and clustering algorithm in the verbal classroom interaction behavior analysis task is improved, and a frame feature extraction network integrating LSTM and TDNN and a temporal pooling network based on the dual multi-head attention mechanism are proposed. Compared with the DIHARD III baseline network, the improved speaker segmentation clustering algorithm reduces the speaker separation error rate (DER) by 3.24%, 3.19%, 4.53% and 4.14%, respectively, on the four types of evaluation datasets. (3) For the face detection algorithm in the non-verbal classroom interactive behavior analysis task, a single-stage face detection network FDN is proposed, and a bidirectional feature fusion module FPN+PANet, a prediction branch IoU- aware and a loss function CIoU are designed. Compared with RetinaFace, the final FDN has the most obvious improvement, and the average precision (Average Precision, AP) on the verification and test set difficult targets has increased by 2.6% and 2.7%, respectively.","PeriodicalId":348024,"journal":{"name":"Proceedings of the 5th International Conference on Control and Computer Vision","volume":"22 11 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134387920","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"A Robust Image Watermarking Algorithm Based on Content Authentication and Intelligent Optimization","authors":"Yi-Lin Bei, Xiaorong Zhu, Qian Zhang, Sai Qiao","doi":"10.1145/3561613.3561639","DOIUrl":"https://doi.org/10.1145/3561613.3561639","url":null,"abstract":"Aiming at some shortcomings of current digital watermarking algorithms, a color image double watermarking algorithm based on content authentication and machine learning is proposed in this paper. Firstly, the algorithm combines discrete wavelet transform (DWT) and discrete cosine transform (DCT), and embeds the watermark through singular value decomposition (SVD) modulation. Secondly, using the learning and classification characteristics of Support Vector Machine (SVM), the watermark detection model is obtained through training a large number of data, and finally the robust watermark is extracted automatically. In order to realize content authentication and tamper location, the algorithm embeds a fragile watermark at the same time, and uses the relationship between the high-frequency coefficients of wavelet transform to embed and extract the watermark. Through the simulation results and data analysis, the proposed double watermarking algorithm not only realizes the intelligent watermark extraction process, but also can accurately locate malicious tampering while maintaining strong robustness.","PeriodicalId":348024,"journal":{"name":"Proceedings of the 5th International Conference on Control and Computer Vision","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131191556","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"An Adaptive Population-based Iterative Greedy Algorithm for Optimizing the Maximum Completion Time of Hybrid Flow Shop","authors":"Fuyou Mao, Xiyang Liu, Haomin Zhao","doi":"10.1145/3561613.3561642","DOIUrl":"https://doi.org/10.1145/3561613.3561642","url":null,"abstract":"Hybrid flow-shop scheduling problem, HFSP is the most common scheduling problem in actual production, the improvement and innovation of its intelligent optimization algorithm has important research value and practical significance. In this paper, we propose an adaptive population-based iterated greedy algorithm (SIGA) to solve the objective function of maximum completion time in production scheduling. Firstly, the NEH (Nawaz-Enscore-Ham) algorithm is used to improve the quality of the initial population; secondly, the destruction and construction operations of the population iterative greedy algorithm are applied to further optimize the population and use the disturbance factor to achieve the adaptive nature of the algorithm to the arithmetic cases; finally, an optimization rate of 86.6% is experimentally derived to obtain a smaller maximum completion time.","PeriodicalId":348024,"journal":{"name":"Proceedings of the 5th International Conference on Control and Computer Vision","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128280718","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Dengfeng Sha, Yuhui Ma, Dan Zhang, Jiong Zhang, Yitian Zhao
{"title":"DeLashNet: A Deep Network for Eyelash Artifact Removal in Ultra-Wide-Field Fundus Images","authors":"Dengfeng Sha, Yuhui Ma, Dan Zhang, Jiong Zhang, Yitian Zhao","doi":"10.1145/3561613.3561649","DOIUrl":"https://doi.org/10.1145/3561613.3561649","url":null,"abstract":"The interference of eyelash artifacts in ultra-wide-field fundus (UWF) images has always been a serious problem in preventing precise clinical observations of pathology. Currently, the automatic removal of eyelash artifacts in UWF images remains unsolved and thus will eventually affect the diagnosis accuracy. In this paper, we propose a deep learning architecture called DeLashNet to eliminate eyelash artifacts from UWF images. Our DeLashNet consists of two stages: the first stage is the eyelash artifact removal stage based on a conditional generative adversarial network, and the second stage is the background refinement stage using an encoder-decoder structure. To solve the issue of lacking training samples with eyelashes, we design a novel eyelash growing model to generate synthetic eyelashes with labels and finally established a paired synthetic eyelashes (PSE) dataset. Experiments are conducted to verify the effectiveness of our proposed DeLashNet on eyelash artifact removal. The comparative and ablation studies demonstrate that the proposed DeLashNet achieved satisfactory removal performance on eyelash artifacts of UWF images.","PeriodicalId":348024,"journal":{"name":"Proceedings of the 5th International Conference on Control and Computer Vision","volume":"30 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127541299","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}