Jinhui Tang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Jinhui Tang has a total of 12 co-authors. Here are the top co-authors by number of shared publications:
Xiuqing Wu
Dept. of Electron. Eng. & Inf. Sci., Univ. of Sci. & Technol. of China, Hefei
3 shared publicationsXin Yang
2 shared publicationsChunyan Liu
1 shared publicationGuo-Jun Qi
Dept. of Autom., Univ. of Sci. & Technol. of China, Hefei
1 shared publication- Guo-Jun Qi
Dept. of EEIS, University of Sci. & Tech. of China, Hefei 230027, P. R. China
1 shared publication - Hong-Jiang Zhang
Microsoft Advanced Technology Center
1 shared publication
Latest Publications
- 16 December 2008
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics)
- Vol. 39 (2) , 409-416
Correlative Linear Neighborhood Propagation for Video Annotation
Recently, graph-based semi-supervised learning methods have been widely applied in multimedia research area. However, for the application of video semantic annotation in multi-label setting, these methods neglect an important characteristic of video data: The semantic concepts appear correlatively and interact naturally with each other rather than exist in isolation. In this paper, we adapt this semantic correlation into graph-based semi-supervised learning and propose a novel method named correlative linear neighborhood propagation to improve annotation performance. Experiments conducted on the Text REtrieval Conference VIDeo retrieval evaluation data set have demonstrated its effectiveness and efficiency.
- 16 December 2008
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics)
- Vol. 39 (2) , 409-416
Correlative Linear Neighborhood Propagation for Video Annotation
Recently, graph-based semi-supervised learning methods have been widely applied in multimedia research area. However, for the application of video semantic annotation in multi-label setting, these methods neglect an important characteristic of video data: The semantic concepts appear correlatively and interact naturally with each other rather than exist in isolation. In this paper, we adapt this semantic correlation into graph-based semi-supervised learning and propose a novel method named correlative linear neighborhood propagation to improve annotation performance. Experiments conducted on the Text REtrieval Conference VIDeo retrieval evaluation data set have demonstrated its effectiveness and efficiency.
- 1 June 2007
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- No. 10636919,p. 1-8
Concurrent Multiple Instance Learning for Image Categorization
We propose a new multiple instance learning (MIL) algorithm to learn image categories. Unlike existing MIL algorithms, in which the individual instances in a bag are assumed to be independent with each other, we develop concurrent tensors to explicitly model the inter-dependency between the instances to better capture image's inherent semantics. Rank-1 tensor factorization is then applied to obtain the label of each instance. Furthermore, we formulate the classification problem in the reproducing kernel Hilbert space (RKHS) to extend instance label prediction to the whole feature space. Finally, a regularizer is introduced, which avoids overfitting and significantly improves learning machine's generalization capability, similar to that in SVMs. We report superior categorization performances compared with key existing approaches on both the COREL and the Caltech datasets.
- 1 June 2007
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- No. 10636919,p. 1-8
Concurrent Multiple Instance Learning for Image Categorization
We propose a new multiple instance learning (MIL) algorithm to learn image categories. Unlike existing MIL algorithms, in which the individual instances in a bag are assumed to be independent with each other, we develop concurrent tensors to explicitly model the inter-dependency between the instances to better capture image's inherent semantics. Rank-1 tensor factorization is then applied to obtain the label of each instance. Furthermore, we formulate the classification problem in the reproducing kernel Hilbert space (RKHS) to extend instance label prediction to the whole feature space. Finally, a regularizer is introduced, which avoids overfitting and significantly improves learning machine's generalization capability, similar to that in SVMs. We report superior categorization performances compared with key existing approaches on both the COREL and the Caltech datasets.
- 1 January 2006
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 2, 9813-9816
Image Denoising Based on Edge Recovery
Common smoothing algorithms can be considered as different kinds of diffusion, which try to make a trade off between denoising and edge information reservation. In this paper, however, we deal with the denoising problem from a novel thought-employing an inverse diffusion to recover edge information after image smoothing, combining adaptive scale filtering and inverse diffusion function for image denoising. This method adopts minimal description length (MDL) rule to adaptively choose the optimal scale for each pixel to smooth the image and improves inverse diffusion function to recover the edges in the degraded image. Experiments show that this method is more effective than classical filters and anisotropic diffusion
- 1 January 2006
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 2, 9813-9816
Image Denoising Based on Edge Recovery
Common smoothing algorithms can be considered as different kinds of diffusion, which try to make a trade off between denoising and edge information reservation. In this paper, however, we deal with the denoising problem from a novel thought-employing an inverse diffusion to recover edge information after image smoothing, combining adaptive scale filtering and inverse diffusion function for image denoising. This method adopts minimal description length (MDL) rule to adaptively choose the optimal scale for each pixel to smooth the image and improves inverse diffusion function to recover the edges in the degraded image. Experiments show that this method is more effective than classical filters and anisotropic diffusion
- 1 January 2006
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 2, 9782-9785
Image Registration Based on Fitted Straight Lines of Edges
Rotation problem is a difficulty in common image registration methods for computational complexity. Aiming at this, a novel method based on fitted straight lines of edges is proposed. It first estimates the rotating angle by angle difference histogram of fitted straight lines, and then estimates the translation by Hausdorff distance. After compensating the rotation and translation, mutual information based on Renyi entropy is used to find the accurate registration result in a small domain. Experiment results demonstrate that this method is effective in solving rotation problem
- 1 January 2006
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 2, 9782-9785
Image Registration Based on Fitted Straight Lines of Edges
Rotation problem is a difficulty in common image registration methods for computational complexity. Aiming at this, a novel method based on fitted straight lines of edges is proposed. It first estimates the rotating angle by angle difference histogram of fitted straight lines, and then estimates the translation by Hausdorff distance. After compensating the rotation and translation, mutual information based on Renyi entropy is used to find the accurate registration result in a small domain. Experiment results demonstrate that this method is effective in solving rotation problem