2017 IEEE International Conference on Image Processing (ICIP)最新文献_第6页

Data-driven assimilation of irregularly-sampled image time series 不规则采样图像时间序列的数据驱动同化

2017 IEEE International Conference on Image Processing (ICIP) Pub Date : 2017-09-17 DOI: 10.1109/ICIP.2017.8297094

Ronan Fablet, Phi Huynh Viet, Redouane Lguensat, B. Chapron

引用次数: 4

An adaptive perceptual quantization method for HDR video coding 一种HDR视频编码的自适应感知量化方法

2017 IEEE International Conference on Image Processing (ICIP) Pub Date : 2017-09-17 DOI: 10.1109/ICIP.2017.8296437

Yi Liu, N. Sidaty, W. Hamidouche, O. Déforges, G. Valenzise, Emin Zerman

引用次数: 8

Transferring CNNS to multi-instance multi-label classification on small datasets 将CNNS转换为小数据集上的多实例多标签分类

2017 IEEE International Conference on Image Processing (ICIP) Pub Date : 2017-09-17 DOI: 10.1109/ICIP.2017.8296498

Mingzhi Dong, Kunkun Pang, Yang Wu, Jing-Hao Xue, Timothy M. Hospedales, T. Ogasawara

{"title":"Transferring CNNS to multi-instance multi-label classification on small datasets","authors":"Mingzhi Dong, Kunkun Pang, Yang Wu, Jing-Hao Xue, Timothy M. Hospedales, T. Ogasawara","doi":"10.1109/ICIP.2017.8296498","DOIUrl":"https://doi.org/10.1109/ICIP.2017.8296498","url":null,"abstract":"Image tagging is a well known challenge in image processing. It is typically addressed through multi-instance multi-label (MIML) classification methodologies. Convolutional Neural Networks (CNNs) possess great potential to perform well on MIML tasks, since multi-level convolution and max pooling coincide with the multi-instance setting and the sharing of hidden representation may benefit multi-label modeling. However, CNNs usually require a large amount of carefully labeled data for training, which is hard to obtain in many real applications. In this paper, we propose a new approach for transferring pre-trained deep networks such as VGG16 on Imagenet to small MIML tasks. We extract features from each group of the network layers and apply multiple binary classifiers to them for multi-label prediction. Moreover, we adopt an L1-norm regularized Logistic Regression (L1LR) to find the most effective features for learning the multi-label classifiers. The experiment results on two most-widely used and relatively small benchmark MIML image datasets demonstrate that the proposed approach can substantially outperform the state-of-the-art algorithms, in terms of all popular performance metrics.","PeriodicalId":229602,"journal":{"name":"2017 IEEE International Conference on Image Processing (ICIP)","volume":"14 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-09-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128832201","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: 4

A data-driven approach to feature space selection for robust micro-endoscopic image reconstruction 基于数据驱动的微内窥镜图像鲁棒重建特征空间选择方法

2017 IEEE International Conference on Image Processing (ICIP) Pub Date : 2017-09-17 DOI: 10.1109/ICIP.2017.8296680

P. Bourdon, D. Helbert

引用次数: 3

Performance comparison of Bayesian iterative algorithms for three classes of sparsity enforcing priors with application in computed tomography 贝叶斯迭代算法在三类稀疏增强先验的性能比较及其在计算机断层扫描中的应用

2017 IEEE International Conference on Image Processing (ICIP) Pub Date : 2017-09-17 DOI: 10.1109/ICIP.2017.8296949

Mircea Dumitru, Li Wang, N. Gac, A. Mohammad-Djafari

引用次数: 5

The Wits Intelligent Teaching System: Detecting student engagement during lectures using convolutional neural networks Wits智能教学系统:使用卷积神经网络检测学生在讲座中的参与度

2017 IEEE International Conference on Image Processing (ICIP) Pub Date : 2017-09-17 DOI: 10.1109/ICIP.2017.8296804

Richard Klein, T. Çelik

引用次数: 19

Unsupervised hyperspectral band selection via multi-feature information-maximization clustering 基于多特征信息最大化聚类的无监督高光谱波段选择

2017 IEEE International Conference on Image Processing (ICIP) Pub Date : 2017-09-17 DOI: 10.1109/ICIP.2017.8296339

M. Bevilacqua, Y. Berthoumieu

引用次数: 12

Deep regional feature pooling for video matching 基于深度区域特征池的视频匹配

2017 IEEE International Conference on Image Processing (ICIP) Pub Date : 2017-09-17 DOI: 10.1109/ICIP.2017.8296307

Yan Bai, Jie Lin, V. Chandrasekhar, Yihang Lou, Shiqi Wang, Ling-yu Duan, Tiejun Huang, A. Kot

引用次数: 1

Deep discovery of facial motions using a shallow embedding layer 使用浅嵌入层深度发现面部运动

2017 IEEE International Conference on Image Processing (ICIP) Pub Date : 2017-09-17 DOI: 10.1109/ICIP.2017.8296545

Afsane Ghasemi, Mahsa Baktash, S. Denman, S. Sridharan, Dung Nguyen Tien, C. Fookes

{"title":"Deep discovery of facial motions using a shallow embedding layer","authors":"Afsane Ghasemi, Mahsa Baktash, S. Denman, S. Sridharan, Dung Nguyen Tien, C. Fookes","doi":"10.1109/ICIP.2017.8296545","DOIUrl":"https://doi.org/10.1109/ICIP.2017.8296545","url":null,"abstract":"Unique encoding of the dynamics of facial actions has potential to provide a spontaneous facial expression recognition system. The most promising existing approaches rely on deep learning of facial actions. However, current approaches are often computationally intensive and require a great deal of memory/processing time, and typically the temporal aspect of facial actions are often ignored, despite the potential wealth of information available from the spatial dynamic movements and their temporal evolution over time from neutral state to apex state. To tackle aforementioned challenges, we propose a deep learning framework by using the 3D convolutional filters to extract spatio-temporal features, followed by the LSTM network which is able to integrate the dynamic evolution of short-duration of spatio-temporal features as an emotion progresses from the neutral state to the apex state. In order to reduce the redundancy of parameters and accelerate the learning of the recurrent neural network, we propose a shallow embedding layer to reduce the number of parameters in the LSTM by up to 98% without sacrificing recognition accuracy. As the fully connected layer approximately contains 95% of the parameters in the network, we decrease the number of parameters in this layer before passing features to the LSTM network, which significantly improves training speed and enables the possibility of deploying a state of the art deep network on real-time applications. We evaluate our proposed framework on the DISFA and UNBC-McMaster Shoulder pain datasets.","PeriodicalId":229602,"journal":{"name":"2017 IEEE International Conference on Image Processing (ICIP)","volume":"836 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-09-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116423399","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: 4

Reflectance-based surface saliency 基于反射率的表面显著性

2017 IEEE International Conference on Image Processing (ICIP) Pub Date : 2017-09-17 DOI: 10.1109/ICIP.2017.8296320

Gilles Pitard, Gaëtan Le Goïc, A. Mansouri, H. Favrelière, Maurice Pillet, S. George, J. Hardeberg

引用次数: 13