Chunheng Wang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Chunheng Wang has a total of 26 co-authors. Here are the top co-authors by number of shared publications:
- Baihua Xiao
State Key Laboratory for Management and Control of Complex Systems, Institute of Automation, Chinese Academy of Sciences, Beijing, China
11 shared publications Baihua Xiao
State Key Laboratory of Management and Control for Complex Systems, Institute of Automation, Chinese Academy of Sciences, Beijing, China
6 shared publications- Cunzhao Shi
Institute of Automation Chinese Academy of Sciences(CASIA)
6 shared publications - Chengzuo Qi
Institute of Automation, Chinese Academy of Sciences, Institute of Automation, Chinese Academy of Sciences, Beijing China
4 shared publications Cunzhao Shi
State Key Laboratory of Management and Control for Complex Systems, Institute of Automation, Chinese Academy of Sciences, Beijing, China
3 shared publications- Yanna Wang
The Key Laboratory of Cognition and Decision Intelligence for Complex Systems Institute of Automation Chinese Academy of Sciences
3 shared publications
Latest Publications
- Open Access
Selective Feature Connection Mechanism: Concatenating Multi-layer CNN Features with a Feature Selector
Different layers of deep convolutional neural networks(CNNs) can encode different-level information. High-layer features always contain more semantic information, and low-layer features contain more detail information. However, low-layer features suffer from the background clutter and semantic ambiguity. During visual recognition, the feature combination of the low-layer and high-level features plays an important role in context modulation. If directly combining the high-layer and low-layer features, the background clutter and semantic ambiguity may be caused due to the introduction of detailed information. In this paper, we propose a general network architecture to concatenate CNN features of different layers in a simple and effective way, called Selective Feature Connection Mechanism (SFCM). Low-level features are selectively linked to high-level features with a feature selector which is generated by high-level features. The proposed connection mechanism can effectively overcome the above-mentioned drawbacks. We demonstrate the effectiveness, superiority, and universal applicability of this method on multiple challenging computer vision tasks, including image classification, scene text detection, and image-to-image translation.
Open AccessSelective Feature Connection Mechanism: Concatenating Multi-layer CNN Features with a Feature Selector
Different layers of deep convolutional neural networks(CNNs) can encode different-level information. High-layer features always contain more semantic information, and low-layer features contain more detail information. However, low-layer features suffer from the background clutter and semantic ambiguity. During visual recognition, the feature combination of the low-layer and high-level features plays an important role in context modulation. If directly combining the high-layer and low-layer features, the background clutter and semantic ambiguity may be caused due to the introduction of detailed information. In this paper, we propose a general network architecture to concatenate CNN features of different layers in a simple and effective way, called Selective Feature Connection Mechanism (SFCM). Low-level features are selectively linked to high-level features with a feature selector which is generated by high-level features. The proposed connection mechanism can effectively overcome the above-mentioned drawbacks. We demonstrate the effectiveness, superiority, and universal applicability of this method on multiple challenging computer vision tasks, including image classification, scene text detection, and image-to-image translation.
- 6 December 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Multimedia
- Vol. 21 (6) , 1551-1562
Iterative Manifold Embedding Layer Learned by Incomplete Data for Large-Scale Image Retrieval
Existing manifold learning methods are not appropriate for image retrieval task, because most of them are unable to process query image and they have much additional computational cost especially for large scale database. Therefore, we propose the iterative manifold embedding (IME) layer, of which the weights are learned off-line by unsupervised strategy, to explore the intrinsic manifolds by incomplete data. On the large scale database that contains 27000 images, IME layer is more than 120 times faster than other manifold learning methods to embed the original representations at query time. We embed the original descriptors of database images which lie on manifold in a high dimensional space into manifold-based representations iteratively to generate the IME representations in off-line learning stage. According to the original descriptors and the IME representations of database images, we estimate the weights of IME layer by ridge regression. In on-line retrieval stage, we employ the IME layer to map the original representation of query image with ignorable time cost (2 milliseconds per image). We experiment on five public standard datasets for image retrieval. The proposed IME layer significantly outperforms related dimension reduction methods and manifold learning methods. Without post-processing, Our IME layer achieves a boost in performance of state-of-the-art image retrieval methods with post-processing on most datasets, and needs less computational cost. Code is available at https://github.com/XJhaoren/IME_layer.
- 6 December 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Multimedia
- Vol. 21 (6) , 1551-1562
Iterative Manifold Embedding Layer Learned by Incomplete Data for Large-Scale Image Retrieval
Existing manifold learning methods are not appropriate for image retrieval task, because most of them are unable to process query image and they have much additional computational cost especially for large scale database. Therefore, we propose the iterative manifold embedding (IME) layer, of which the weights are learned off-line by unsupervised strategy, to explore the intrinsic manifolds by incomplete data. On the large scale database that contains 27000 images, IME layer is more than 120 times faster than other manifold learning methods to embed the original representations at query time. We embed the original descriptors of database images which lie on manifold in a high dimensional space into manifold-based representations iteratively to generate the IME representations in off-line learning stage. According to the original descriptors and the IME representations of database images, we estimate the weights of IME layer by ridge regression. In on-line retrieval stage, we employ the IME layer to map the original representation of query image with ignorable time cost (2 milliseconds per image). We experiment on five public standard datasets for image retrieval. The proposed IME layer significantly outperforms related dimension reduction methods and manifold learning methods. Without post-processing, Our IME layer achieves a boost in performance of state-of-the-art image retrieval methods with post-processing on most datasets, and needs less computational cost. Code is available at https://github.com/XJhaoren/IME_layer.
- 18 September 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 57 (3) , 1358-1367
A Selection Criterion for the Optimal Resolution of Ground-Based Remote Sensing Cloud Images for Cloud Classification
In ground-based remote sensing cloud image observation, images with the highest possible resolution are captured to obtain sufficient information about clouds. However, when features are extracted and classification is performed on the basis of the original images, a high-resolution probably means a high (or even more, unacceptable) computation cost. In practical application, a simple and commonly adopted method is to appropriately resize the original image to a version with a decreased resolution. An inevitable problem is whether useful information is lost in this resizing operation. This paper demonstrates that information loss is inevitable and poor classification results may be obtained from the analysis of local binary pattern (LBP) histogram features. However, this problem has been always neglected in previous studies, and the original image is arbitrarily resized without any criterion. In particular, the histogram features based on LBPs actually reflect the distribution of features. Thus, a criterion based on the Kullback-Leibler divergence between LBP histograms from the original and resized images and a penalty term imposed on the resolution are proposed to select the resolution of the resized image. The optimal resolution of the resized image can be selected by minimizing this criterion. Furthermore, experiments based on three ground-based remote sensing cloud image data sets with different original resolutions validate this criterion by analyzing the LBP histogram features.
- 18 September 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 57 (3) , 1358-1367
A Selection Criterion for the Optimal Resolution of Ground-Based Remote Sensing Cloud Images for Cloud Classification
In ground-based remote sensing cloud image observation, images with the highest possible resolution are captured to obtain sufficient information about clouds. However, when features are extracted and classification is performed on the basis of the original images, a high-resolution probably means a high (or even more, unacceptable) computation cost. In practical application, a simple and commonly adopted method is to appropriately resize the original image to a version with a decreased resolution. An inevitable problem is whether useful information is lost in this resizing operation. This paper demonstrates that information loss is inevitable and poor classification results may be obtained from the analysis of local binary pattern (LBP) histogram features. However, this problem has been always neglected in previous studies, and the original image is arbitrarily resized without any criterion. In particular, the histogram features based on LBPs actually reflect the distribution of features. Thus, a criterion based on the Kullback-Leibler divergence between LBP histograms from the original and resized images and a penalty term imposed on the resolution are proposed to select the resolution of the resized image. The optimal resolution of the resized image can be selected by minimizing this criterion. Furthermore, experiments based on three ground-based remote sensing cloud image data sets with different original resolutions validate this criterion by analyzing the LBP histogram features.
- 1 June 2018
- Published by Elsevier in Neurocomputing
- Vol. 295, 46-58
CRF based text detection for natural scene images using convolutional neural network and context information
- 1 June 2018
- Published by Elsevier in Neurocomputing
- Vol. 295, 46-58
CRF based text detection for natural scene images using convolutional neural network and context information
- 19 December 2017
- Published by Taylor & Francis in Remote Sensing Letters
- Vol. 9 (3) , 274-283
Short-term cloud coverage prediction using the ARIMA time series model
In view of the important role of cloud coverage on the solar (energy) irradiance, the total cloud coverage prediction based on ground-based cloud images is studied in this paper. In traditional prediction techniques, the correlation between cloud coverage over continue time is always neglected. Thus, an autoregressive integrated moving average (ARIMA) time series model is used to predict the short-term cloud coverage. Experimental results on a collected time series database of cloud coverage computed from ground-based cloud images show that the correlation information of time series is useful for cloud coverage prediction. Additionally, the ARIMA model gains a superior prediction performance for forecasts of one minute or longer 20 and 30 minutes. We are able to predict the cloud coverage with an approximate error of 5%, 7%, and 9% for 1, 5, and 20 and 30 minute forecasts, respectively. Furthermore, we found that there are different error rates of predictions for different cloud coverage intervals. High cloud coverage always suffers from a higher error rate.
- 19 December 2017
- Published by Taylor & Francis in Remote Sensing Letters
- Vol. 9 (3) , 274-283
Short-term cloud coverage prediction using the ARIMA time series model
In view of the important role of cloud coverage on the solar (energy) irradiance, the total cloud coverage prediction based on ground-based cloud images is studied in this paper. In traditional prediction techniques, the correlation between cloud coverage over continue time is always neglected. Thus, an autoregressive integrated moving average (ARIMA) time series model is used to predict the short-term cloud coverage. Experimental results on a collected time series database of cloud coverage computed from ground-based cloud images show that the correlation information of time series is useful for cloud coverage prediction. Additionally, the ARIMA model gains a superior prediction performance for forecasts of one minute or longer 20 and 30 minutes. We are able to predict the cloud coverage with an approximate error of 5%, 7%, and 9% for 1, 5, and 20 and 30 minute forecasts, respectively. Furthermore, we found that there are different error rates of predictions for different cloud coverage intervals. High cloud coverage always suffers from a higher error rate.
- Open Access
Unsupervised Part-based Weighting Aggregation of Deep Convolutional Features for Image Retrieval
In this paper, we propose a simple but effective semantic part-based weighting aggregation (PWA) for image retrieval. The proposed PWA utilizes the discriminative filters of deep convolutional layers as part detectors. Moreover, we propose the effective unsupervised strategy to select some part detectors to generate the "probabilistic proposals", which highlight certain discriminative parts of objects and suppress the noise of background. The final global PWA representation could then be acquired by aggregating the regional representations weighted by the selected "probabilistic proposals" corresponding to various semantic content. We conduct comprehensive experiments on four standard datasets and show that our unsupervised PWA outperforms the state-of-the-art unsupervised and supervised aggregation methods. Code is available at https://github.com/XJhaoren/PWA.
Open AccessUnsupervised Part-based Weighting Aggregation of Deep Convolutional Features for Image Retrieval
In this paper, we propose a simple but effective semantic part-based weighting aggregation (PWA) for image retrieval. The proposed PWA utilizes the discriminative filters of deep convolutional layers as part detectors. Moreover, we propose the effective unsupervised strategy to select some part detectors to generate the "probabilistic proposals", which highlight certain discriminative parts of objects and suppress the noise of background. The final global PWA representation could then be acquired by aggregating the regional representations weighted by the selected "probabilistic proposals" corresponding to various semantic content. We conduct comprehensive experiments on four standard datasets and show that our unsupervised PWA outperforms the state-of-the-art unsupervised and supervised aggregation methods. Code is available at https://github.com/XJhaoren/PWA.
- 1 October 2017
- Published by Elsevier in Pattern Recognition Letters
- Vol. 97, 69-76
Multi-order co-occurrence activations encoded with Fisher Vector for scene character recognition
- 1 October 2017
- Published by Elsevier in Pattern Recognition Letters
- Vol. 97, 69-76
Multi-order co-occurrence activations encoded with Fisher Vector for scene character recognition
- Open Access
- 1 August 2015
- Published by Institution of Engineering and Technology (IET) in IET Computer Vision
- Vol. 9 (4) , 531-540
Cross‐view face recognition via structured dictionary based domain shift
View variation is a major challenge in face recognition. In this study, the authors propose a novel cross‐view face recognition method by seeking potential intermediate domains between the source and target views to model the connection of varying‐views faces. Specifically, each intermediate domain is associated with a dictionary subspace. Learning proceeds in two phases. First, the authors discriminatively train a sub‐dictionary for each subclass of data, which then compose a structured dictionary of powerful reconstructive and discriminative capability on the source data. Secondly, the authors gradually adapt the source domain dictionary to the target domain by incrementally reducing the reconstruction error on the target data, which forms a smooth transition path connecting the source and target domains. Instead of updating the structured dictionary integrally, the authors develop a refined sub‐dictionary‐based updating algorithm, which makes the intermediate dictionaries fit on the target data better and faster. Finally, the authors apply invariant sparse codes across the source, intermediate and target domains to render domain‐shared representations, where the sample differences caused by view changes are reduced. Experiments on the CMU‐PIE and Multi‐PIE dataset demonstrate the effectiveness of the proposed method.
- 1 August 2015
- Published by Institution of Engineering and Technology (IET) in IET Computer Vision
- Vol. 9 (4) , 531-540
Open AccessCross‐view face recognition via structured dictionary based domain shift
View variation is a major challenge in face recognition. In this study, the authors propose a novel cross‐view face recognition method by seeking potential intermediate domains between the source and target views to model the connection of varying‐views faces. Specifically, each intermediate domain is associated with a dictionary subspace. Learning proceeds in two phases. First, the authors discriminatively train a sub‐dictionary for each subclass of data, which then compose a structured dictionary of powerful reconstructive and discriminative capability on the source data. Secondly, the authors gradually adapt the source domain dictionary to the target domain by incrementally reducing the reconstruction error on the target data, which forms a smooth transition path connecting the source and target domains. Instead of updating the structured dictionary integrally, the authors develop a refined sub‐dictionary‐based updating algorithm, which makes the intermediate dictionaries fit on the target data better and faster. Finally, the authors apply invariant sparse codes across the source, intermediate and target domains to render domain‐shared representations, where the sample differences caused by view changes are reduced. Experiments on the CMU‐PIE and Multi‐PIE dataset demonstrate the effectiveness of the proposed method.
- Open Access
- 1 February 2015
- Published by Institution of Engineering and Technology (IET) in IET Computer Vision
- Vol. 9 (1) , 138-148
Scene text recognition by learning co‐occurrence of strokes based on spatiality embedded dictionary
Text information contained in scene images is very helpful for high‐level image understanding. In this study, the authors propose to learn co‐occurrence of local strokes for scene text recognition by using a spatiality embedded dictionary (SED). Unlike spatial pyramid partitioning images into grids to incorporate spatial information, the authors SED associates every codeword with a particular response region and introduces more precise spatial information for robust character recognition. After localised soft coding and max pooling of the first layer, a sparse dictionary is learned to model co‐occurrence of several local strokes, which further improves classification performance. Experimental results on two scene character recognition datasets ICDAR2003 and CHARS74 K demonstrate that their character recognition method outperforms state‐of‐the‐art methods. Besides, competitive word recognition results are also reported for four benchmark word recognition datasets ICDAR2003, ICDAR2011, ICDAR2013 and street view text when combining their character recognition method with a conditional random field language model.
- 1 February 2015
- Published by Institution of Engineering and Technology (IET) in IET Computer Vision
- Vol. 9 (1) , 138-148
Open AccessScene text recognition by learning co‐occurrence of strokes based on spatiality embedded dictionary
Text information contained in scene images is very helpful for high‐level image understanding. In this study, the authors propose to learn co‐occurrence of local strokes for scene text recognition by using a spatiality embedded dictionary (SED). Unlike spatial pyramid partitioning images into grids to incorporate spatial information, the authors SED associates every codeword with a particular response region and introduces more precise spatial information for robust character recognition. After localised soft coding and max pooling of the first layer, a sparse dictionary is learned to model co‐occurrence of several local strokes, which further improves classification performance. Experimental results on two scene character recognition datasets ICDAR2003 and CHARS74 K demonstrate that their character recognition method outperforms state‐of‐the‐art methods. Besides, competitive word recognition results are also reported for four benchmark word recognition datasets ICDAR2003, ICDAR2011, ICDAR2013 and street view text when combining their character recognition method with a conditional random field language model.
- Open Access
- 1 December 2014
- Published by Institution of Engineering and Technology (IET) in IET Computer Vision
- Vol. 8 (6) , 579-587
Human action recognition using weighted pooling
Pooling strategies, such as max pooling and sum pooling, have been widely used to obtain the global representations for action videos. However, these pooling strategies have several disadvantages. First, they are easily affected by unwanted background local features, the absence of discriminative local features and the times of actions periodically performed by actors. Second, most pooling strategies only use local features to build the global representation that captures little mid‐level features for action representation. In this study, the authors propose a novel weighted pooling strategy based on actionlets representation for action recognition. The actionlets are defined as the movements of large bodies such as legs, arms and head, which capture rich mid‐level features for action representation. Besides, the authors’ method also incorporates the distribution information of actionlets into pooling procedure. Specifically, a pooling weight, which determines the importance of actionlet on the final video representation, is assigned to each actionlet. To learn the weight, they propose a novel discriminative learning algorithm to capture the discriminative information for pooling operation. They evaluate their weighted pooling on three datasets: KTH actions dataset, UCF sports dataset and Youtube actions dataset. Experimental results show the effectiveness of the proposed method.
- 1 December 2014
- Published by Institution of Engineering and Technology (IET) in IET Computer Vision
- Vol. 8 (6) , 579-587
Open AccessHuman action recognition using weighted pooling
Pooling strategies, such as max pooling and sum pooling, have been widely used to obtain the global representations for action videos. However, these pooling strategies have several disadvantages. First, they are easily affected by unwanted background local features, the absence of discriminative local features and the times of actions periodically performed by actors. Second, most pooling strategies only use local features to build the global representation that captures little mid‐level features for action representation. In this study, the authors propose a novel weighted pooling strategy based on actionlets representation for action recognition. The actionlets are defined as the movements of large bodies such as legs, arms and head, which capture rich mid‐level features for action representation. Besides, the authors’ method also incorporates the distribution information of actionlets into pooling procedure. Specifically, a pooling weight, which determines the importance of actionlet on the final video representation, is assigned to each actionlet. To learn the weight, they propose a novel discriminative learning algorithm to capture the discriminative information for pooling operation. They evaluate their weighted pooling on three datasets: KTH actions dataset, UCF sports dataset and Youtube actions dataset. Experimental results show the effectiveness of the proposed method.