International Conference on Image and Graphics Processing最新文献

Densely connected AutoEncoders for image compression 密集连接的自动编码器用于图像压缩

International Conference on Image and Graphics Processing Pub Date : 1900-01-01 DOI: 10.1145/3313950.3313965

Zebang Song, S. Kamata

{"title":"Densely connected AutoEncoders for image compression","authors":"Zebang Song, S. Kamata","doi":"10.1145/3313950.3313965","DOIUrl":"https://doi.org/10.1145/3313950.3313965","url":null,"abstract":"Image compression, which is a type of data compression applied to digital images, has been a fundamental research topic for many decades. Recent image techniques produce very large amounts of data, which may make it prohibitive to storage and communications of image data without the use of compression. However, the traditional compression methods, such as JPEG, may introduce the compression artefact problems. Recently, deep learning has achieved great success in many computer vision tasks and is gradually being used in image compression. To solve the compression atrefact problem, in this paper, we present a lossy image compression architecture, which utilizes the advantages of the existing deep learning methods to achieve a high coding efficiency. We design a densely connected autoencoder structure for lossy image compression. Firstly, we design a densely autoencoder structure to get richer feature information from image which can be helpful for compression. Secondly, we design a U-net like network to decrease the distortion caused by compression. Finally, an improved binarizer is adopted to quantize the output of encoder. In low bit rate image compression, experiments show that our method significantly outperforms JPEG and JPEG2000 and can produce a better visual result with sharp edges, rich textures, and fewer artifacts.","PeriodicalId":149627,"journal":{"name":"International Conference on Image and Graphics Processing","volume":"75 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1900-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115666913","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}

引用次数: 8

Pedestrian Behavior Anomaly Detection Based on Dynamic Mode Decomposition and One-Class SVM 基于动态模式分解和一类支持向量机的行人行为异常检测

International Conference on Image and Graphics Processing Pub Date : 1900-01-01 DOI: 10.1145/3447587.3447589

Weixi Zhang, Shuai Dong, Kun Zou, Wensheng Li

{"title":"Pedestrian Behavior Anomaly Detection Based on Dynamic Mode Decomposition and One-Class SVM","authors":"Weixi Zhang, Shuai Dong, Kun Zou, Wensheng Li","doi":"10.1145/3447587.3447589","DOIUrl":"https://doi.org/10.1145/3447587.3447589","url":null,"abstract":"Anomaly detection of human behavior, which has been widely applied in video supervision in recent years, aims at recognizing the human behaviors that are out of the normal scope in high-level semantic. With the development of deep learning (DL), many DL-based methods for behavior recognition have been developed. However, the lack of negative samples hinders the application of DL-based methods. Thus, this work proposed a new framework based on dynamic mode decomposition (DMD)which is free from the requirements of negative samples. First, key points of the human body in each frame of a video are exacted by the -deep neural networks (DNN)which have been solved in recent studies. Second, the key points are projected into a mode space of much lower dimensions with DMD. At last, the detection results are obtained with One-class SVM, which is a classification method that does not require negative samples. The data set we use in the experiment is the CMU motion data set. Experiment show that the accuracy of DMD combined with support vector classification (SVC) achieve 85% while the accuracy of DMD combined with one-class SVM achieve 81%, and the proposed framework can distinguish anomaly behavior with a small number of samples.","PeriodicalId":149627,"journal":{"name":"International Conference on Image and Graphics Processing","volume":"130 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1900-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116207269","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}

引用次数: 0

MGF: An Algorithm for Compressed Sensing MRI with Gradient Domain Guided Image Filtering 基于梯度域引导图像滤波的压缩感知MRI算法

International Conference on Image and Graphics Processing Pub Date : 1900-01-01 DOI: 10.1145/3447587.3447619

Peixian Zhuang

引用次数: 0

Accurate Thoracic Disease Classification via Ensemble Networks 基于集成网络的胸部疾病准确分类

International Conference on Image and Graphics Processing Pub Date : 1900-01-01 DOI: 10.1145/3512388.3512417

Arren Matthew C. Antioquia

引用次数: 1

Aircraft Detection of High-Resolution Remote Sensing Image Based on Faster R-CNN Model and SSD Model 基于Faster R-CNN模型和SSD模型的高分辨率遥感图像飞机检测

International Conference on Image and Graphics Processing Pub Date : 1900-01-01 DOI: 10.1145/3191442.3191443

Qifang Xie, Guoqing Yao, Ping Liu

{"title":"Aircraft Detection of High-Resolution Remote Sensing Image Based on Faster R-CNN Model and SSD Model","authors":"Qifang Xie, Guoqing Yao, Ping Liu","doi":"10.1145/3191442.3191443","DOIUrl":"https://doi.org/10.1145/3191442.3191443","url":null,"abstract":"With the continuous improvement of the space resolution in remote sensing images, the rapid and accurate detection in high-resolution remote sensing images has become a hotspot in the field of remote sensing application. For nearly 10 years, deep learning has made outstanding achievements in the feature extraction of original image and received attention of a large number of scholars. Among them, the convolutional neural network (CNN) has made breakthrough progress in the field of image classification and detection, and has overcome three shortcomings of the original remote sensing image detection method: low detection efficiency, redundant human resource input, and flawed feature selection. In this paper, Faster R-CNN model and SSD model are trained by high-resolution remote sensing images. The appropriate training time is determined by the detection results of verification set and the loss function. When we get trained models, it will be used to detect the test set images, and the accuracy rate and recall rate of two models were calculated by visual interpretation method. The experimental results show that both the Faster R-CNN model and the SSD model can be applied to aircraft detection in corresponding high-resolution remote sensing images. The SSD model can detect the single scene aircraft quickly and accurately. The Faster R-CNN model has a high accuracy but cannot reach the requirement of real-time detection. Besides, the accuracy rate and recall rate of Faster R-CNN model was significantly higher than the SSD model in the complex scenes, and the Faster R-CNN model has a great advantage for the detection of small aircraft.","PeriodicalId":149627,"journal":{"name":"International Conference on Image and Graphics Processing","volume":"27 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1900-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125017764","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}

引用次数: 1