Optimization for Computer Vision and Machine Learning

Optimization for Computer Vision and Machine Learning

Details

Many applications in computer vision and machine learning can be directly formulated as a nonlinear optimization problem, which may be inherently non-convex or/and NP-hard discrete (e.g. l0 norm sparse optimization, multiple label MRF learning, representation learning, image segmentation, clustering and classification). This becomes computationally very challenging in algorithm design. In our research, we consider and design efficient and effective convex relaxations (l1  norm relaxation, linear box relaxation, SOCP relaxation, SDP relaxation) or multiple-stage convex optimization algorithms (Dinkelbach algorithm, MPEC optimization, DC programming, greedy basis pursuit) to tackle these problems.  
 

Selected Publications



 
 

 Designing Camera Networks by Convex Quadratic Programming




 
 

 

 L0TV : A New Method for Image Restoration in the Presence of Impulse Noise





​​ 


 

 Dinkelbach NCUT: A Framework for Solving Normalized Cuts Problems with Priors and Convex  Constraints





​​

Publications

  1. Ganzhao Yuan and Bernard Ghanem, “L0TV : A New Method for Image Restoration in the Presence of Impulse Noise”, Conference on Computer Vision and Pattern Recognition (CVPR 2015)
  2. Bernard Ghanem, Yuanhao Cao, Peter Wonka, "Designing Camera Networks by Convex Quadratic Programming", Computer Graphics Forum (Proceedings of Eurographics) 2015
  3. Zhifeng Hao, Ganzhao Yuan, Bernard Ghanem, "BILGO: Bilateral Greedy Optimization for Large Scale Semidefinite Programming", Neurocomputing Journal 2013
  4. Ganzhao Yuan, Zhenjie Zhang*, Bernard Ghanem*, and Zhifeng Hao, "Low-Rank Quadratic Semidefinite Programming", Neurocomputing Journal 2012
  5. Bernard Ghanem and Narendra Ahuja, "A Probabilistic Framework for Discriminative Dictionary Learning", (arXiv e-print 2011 arXiv:1109.2389v1 [cs.CV])
  6. Emre Akbas, Bernard Ghanem, and Narendra Ahuja, "MIS-Boost: Multiple Instance Selection Boosting", IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI 2012) [Under Review], (arXiv e-print 2011 arXiv:1109.2388v1 [cs.LG])
  7. Bernard Ghanem and Narendra Ahuja, "Dinkelbach NCUT: A Framework for Solving Normalized Cuts Problems with Priors and Convex Constraints", International Journal of Computer Vision (IJCV 2010)