Abstract
Image recognition with incomplete data is a well-known hard problem in computer vision and machine learning. This article proposes a novel deep learning technique called Field Effect Bilinear Deep Networks (FEBDN) for this problem. To address the difficulties of recognizing incomplete data, we design a novel second-order deep architecture with the Field Effect Restricted Boltzmann Machine, which models the reliability of the delivered information according to the availability of the features. Based on this new architecture, we propose a new three-stage learning procedure with field effect bilinear initialization, field effect abstraction and estimation, and global fine-tuning with missing features adjustment. By integrating the reliability of features into the new learning procedure, the proposed FEBDN can jointly determine the classification boundary and estimate the missing features. FEBDN has demonstrated impressive performance on recognition and estimation tasks in various standard datasets.
- André Aleman, Koen B. E. Böcker, Ron Hijman, Edward H. F. de Haanb, and René S. Kahna. 2003. Cognitive basis of hallucinations in schizophrenia: Role of top-down information processing. Schizophr. Res. 64, 2--3, 178--185.Google ScholarCross Ref
- Pradeep K. Atrey, M. Anwar Hossain, Abdulmotaleb El Saddik, and Mohan S. Kankanhalli. 2010. Multimodal fusion for multimedia analysis: a survey. Multimedia Syst. 16, 345--379. Google ScholarDigital Library
- Bernhard E. Boser, Isabelle M. Guyon, and Vladimir N. Vapnik. 1992. A training algorithm for optimal margin classifiers. In COLT. ACM, New York, NY, 144--152. Google ScholarDigital Library
- Gal Chechik, Geremy Heitz, Gal Elidan, Pieter Abbeel, and Daphne Koller. 2006. Max-margin classification of incomplete data. In NIPS. Google ScholarDigital Library
- Gal Chechik, Geremy Heitz, Gal Elidan, Pieter Abbeel, and Daphne Koller. 2008. Max-margin classification of data with absent features. J. Mach. Learn. Res. 9, 1--21. Google ScholarDigital Library
- Hao Chen, Dong Ni, Jing Qin, Shengli Li, Xin Yang, Tianfu Wang, and Pheng Ann Heng. 2015. Standard plane localization in fetal ultrasound via domain transferred deep neural networks. JBHI 19, 5, 1627--1636.Google Scholar
- Yanjiao Chen, Kaishun Wu, and Qian Zhang. 2015. From QoS to QoE: A tutorial on video quality assessment. IEEE Commun. Surv. Tutorials 17, 2, 1126--1165.Google ScholarDigital Library
- Uwe Dick, Peter Haider, and Tobias Scheffer. 2008. Learning from incomplete data with infinite imputations. In ICML. Citeseerx, Helsinki, Finland, 232--239. Google ScholarDigital Library
- Huijun Ding, Tan Lee, Ing Yann Soon, Chai Kiat Yeo, Peng Dai, and Guo Dan. 2015. Objective measures for quality assessment of noise-suppressed speech. Speech Commun. 71, 62--73. Google ScholarDigital Library
- Laura Folguera, Jure Zupan, Daniel Cicerone, and Jorge F. Magallanes. 2015. Self-organizing maps for imputation of missing data in incomplete data matrices. Chemometr. Intell. Lab. 143, 146--151.Google ScholarCross Ref
- Geoffrey E. Hinton and Roweis R. Salakhutdinov. 2006. Reducing the dimensionality of data with neural networks. Science 313, 5786, 504--507.Google Scholar
- Oliver Jesorsky, Klaus J. Kirchberg, and Robert Frischholz. 2001. Robust face detection using the Hausdorff distance. In AVBPA. Springer-Verlag, London, UK, 90--95. Google ScholarDigital Library
- Alex Krizhevsky and Geoffrey E. Hinton. 2009. Learning multiple layers of features from tiny images. Technical Report. University of Toronto.Google Scholar
- Alex Krizhevsky, Ilya Sutskever, and Geoffrey Hinton. 2012. ImageNet classification with deep convolutional neural networks. In NIPS.Google Scholar
- Honglak Lee, Roger Grosse, Rajesh Ranganath, and Andrew Y. Ng. 2011. Unsupervised learning of hierarchical representations with convolutional deep belief networks. Commun. ACM. 54, 10, 95--103. Google ScholarDigital Library
- Xuejun Liao, Hui Li, and Lawrence Carin. 2007. Quadratically gated mixture of experts for incomplete data classification. In ICML. ACM, New York, NY, 553--560. Google ScholarDigital Library
- Norbert R. Malik. 1995. Electronic Circuits: Analysis, Simulation, and Design. Prentice-Hall, Upper Saddle River, NJ. Google ScholarDigital Library
- Prabhu Natarajan, Pradeep K. Atrey, and Mohan Kankanhalli. 2015. Multi-camera coordination and control in surveillance systems: a survey. ACM TOMM. 11, 4, Article 57, 30. Google ScholarDigital Library
- Marc’aurelio Ranzato, Joshua M. Susskind, Volodymyr Mnih, and Geoffrey E. Hinton. 2011. On deep generative models with applications to recognition. In CVPR. 2857--2864. Google ScholarDigital Library
- Yann LeCun, Léeon Bottou, Yoshua Bengio, and Patrick Haffner. 1998. Gradient-based learning applied to document recognition. Proceedings of the IEEE 86, 11, 2278--2324.Google ScholarCross Ref
- Yann LeCun, Yoshua Bengio, and Geoffrey E. Hinton. 2015. Deep learning. Nature 521, 436--444.Google ScholarCross Ref
- Kun Li, Jingyu Yang, and Jianmin Jiang. 2015. Nonrigid structure from motion via sparse representation. IEEE Trans. Cybern. 45, 8, 1401--1413.Google ScholarCross Ref
- Archana Purwar and Sandeep Kumar Singh. 2015. Hybrid prediction model with missing value imputation for media data. ESWA. 42, 5621--5631. Google ScholarDigital Library
- Ruslan Salakhutdinov, Andriy Mnih, and Geoffrey Hinton. 2007. Restricted Boltzmann machines for collaborative filtering. In ICML. ACM, New York, NY, 791--798. Google ScholarDigital Library
- Ruslan Salakhutdinov and Geoffrey E. Hinton. 2007. Learning a nonlinear embedding by preserving class neighbourhood structure. In AISTATS. Omnipress, San Juan, Puerto Rico, 412--419.Google Scholar
- Jürgen Schmidhuber. 2014. Deep learning in neural networks. Technical Report, 61, 85--117. Google ScholarDigital Library
- Kihyuk Sohn, Guanyu Zhou, Chansoo Lee, and Honglak Lee. 2013. Learning and selecting features jointly with point-wise gated boltzmann machines. In ICML. Citeseerx, Atlanta, GA, 217--225.Google Scholar
- Charlie Tang and Chris Eliasmith. 2010. Deep networks for robust visual recognition. In ICML. ACM, 1055--1062.Google Scholar
- Neill R. Taylor, Christo Panchev, Matthew Hartley, Stathis Kasderidis, and John G. Taylor. 2006. Occlusion, attention and object representations. In ICANN. Springer-Verlag, Athens, Greece, 592--601. Google ScholarDigital Library
- Jason Weston, Frédéric Ratle, and Ronan Collobert. 2008. Deep learning via semi-supervised embedding. In ICML. Springer, Berlin, 639--655. Google ScholarDigital Library
- David Williams, Xuejun Liao, Ya Xue, Lawrence Carin, and Balaji Krishnapuram. 2007. On classification with incomplete data. IEEE TPAMI. 29, 3, 427--436. Google ScholarDigital Library
- David Williams, Xuejun Liao, Ya Xue, and Lawrence Carin. 2005. Incomplete-data classification using logistic regression. In ICML. ACM, New York, NY, 972--979. Google ScholarDigital Library
- Hao-tian Wu, Jiwu Huang, and Yun-Qing Shi. 2015. A reversible data hiding method with contrast enhancement for medical images. J. Vis. Commun. Image R. 31, 146--153. Google ScholarDigital Library
- Wanmin Wu, Ahsan Arefin, Raoul Rivas, Klara Nahrstedt, and Renata M. Sheppard. 2009. Quality of experience in distributed interactive multimedia environments: toward a theoretical framework. In ACM MM. 1--10. Google ScholarDigital Library
- Xiaoshan Yang, Tianzhu Zhang, and Changsheng Xu. 2015. Boosted multifeature learning for cross-domain transfer. ACM TOMM. Appl. 11, 3, Article 35, 18. Google ScholarDigital Library
- Quanzeng You, Jiebo Luo, Hailin Jin, and Jianchao Yang. 2015. Robust image sentiment analysis using progressively trained and domain transferred deep networks. In AAAI. Google ScholarDigital Library
- Sheng-hua Zhong, Yan Liu, and Yang Liu. 2011. Bilinear deep learning for image classification. In ACMMM. ACM, New York, NY, 343--352. Google ScholarDigital Library
- Sheng-hua Zhong, Yan Liu, Fu-lai Chung, and Gangshan Wu. 2012. Semiconducting bilinear deep learning for incomplete image recognition. In ICMR. ACM, New York, NY, Article 32. Google ScholarDigital Library
- Sheng-hua Zhong, Yan Liu, Bin Li, and Jing Long. 2015. Query-oriented unsupervised multi-document summarization via deep learning model. ESWA. 42, 21, 8146--8155. Google ScholarDigital Library
- Mingyuan Zhou, Haojun Chen, John Paisley, Lu Ren, Lingbo Li, Zhengming Xing, David Dunson, Guillermo Sapiro, and Lawrence Carin. 2012. Nonparametric bayesian dictionary learning for analysis of noisy and incomplete images. TIP. 21, 1, 2012. Google ScholarDigital Library
Index Terms
- Field Effect Deep Networks for Image Recognition with Incomplete Data
Recommendations
Semiconducting bilinear deep learning for incomplete image recognition
ICMR '12: Proceedings of the 2nd ACM International Conference on Multimedia RetrievalImage recognition with incomplete data is a well-known hard problem in multimedia content analysis. This paper proposes a novel deep learning technique called semiconducting bilinear deep belief networks (SBDBN) by referencing human's visual cortex and ...
Unsupervised local deep feature for image recognition
ULDF is proposed to make better use of autoencoder for image recognition. It is performed on local patches rather than whole images, which helps to scale the algorithm to realistic-sized images.Owning to the combination with BoW, it is more robust to ...
Efficient deep feature selection for remote sensing image recognition with fused deep learning architectures
AbstractConvolutional neural networks (CNNs) have recently emerged as a popular topic for machine learning in various academic and industrial fields. It is often an important problem to obtain a dataset with an appropriate size for CNN training. However, ...
Comments