Vittal Premachandran

Brief Bio

Vittal Premachandran is a Postdoctoral Researcher at the Johns Hopkins University. He obtained the B.E. degree in Computer Science from Visveswaraya Technological University, India, in 2009, graduating with a First Class with Distinction. He received his Ph.D degree, also in Computer Science, from Nanyang Technological University, Singapore, in 2014. He has previously worked as a Postdoctoral Research Fellow at University of California, Los Angeles, and a visiting researcher at National University of Singapore.

His research interests span computer vision and machine learning, with a focus on semantic analysis of scenes. His work has received the Best Student Paper award at the IEEE International Conference on Image Processing (ICIP) in 2013.

Talks and Lectures

• Tutorial on Efficient Inference in Gaussian CRFs @ CCVL, Johns Hopkins Unviersity, August, 2016.
• Fully Convolutional Neural Networks - Guest Lecture for 'Deep Learning for Image Understanding' @ Johns Hopkins Unviersity, March, 2016.
• Introduction to Deep Learning and using Caffe - Guest Lecture for 'Deep Learning for Image Understanding' @ Johns Hopkins Unviersity, March, 2016.
• Reinforcement Learning - Guest Lecture for Stat 271: Probabilistic Models of the Visual Cortex, UCLA, November, 2015.
• Introduction to Deep Networks - Guest Lecture for Stat 271: Probabilistic Models of the Visual Cortex, UCLA, November, 2015.
• Introduction to Vision - Guest Lecture for Stat 271: Probabilistic Models of the Visual Cortex, UCLA, September, 2015.
• Introduction to Deep Networks - Stat 161/261: Introduction to Machine Learning, UCLA, June 2015.
• K-Means, GMM and EM - Stat 161/261: Introduction to Machine Learning, UCLA, May 2015.
• Decision Trees and Random Forest - Stat 161/261: Introduction to Machine Learning, UCLA, May 2015.
• Decision Trees - Stat 161/261: Introduction to Machine Learning, UCLA, May 2015.
• Soft Margin Support Vector Machines - Stat 161/261: Introduction to Machine Learning, UCLA, April 2015.
• Support Vector Machines - Stat 161/261: Introduction to Machine Learning, UCLA, April 2015.
• Perceptrons and Kernels - Stat 161/261: Introduction to Machine Learning, UCLA, April 2015.
• Introduction to Machine Learning - Stat 161/261: Introduction to Machine Learning, UCLA, March 2015.
• Hedging on diversity and making better predictions using EMBR - Guest Lecture for Stat 238: Vision as Bayesian Inference course @ UCLA, November, 2014.
• Empirical Minimum Bayes Risk Prediction @ CCVL, UCLA, October, 2014
• Exploiting Shape Properties for Improved Retrieval, Discrimination and Recognition @ Amazon, Seattle, May 2014.
• Exploiting Shape Properties for Improved Retrieval, Disctimination and Recognition @ NTU, Singapore, February, 2014

Publications

Pre-prints

Deep Residual Learning for Instrument Segmentation in Robotic Surgery
Daniil Pakhomov, Vittal Premachandran, Max Allan, Mahdi Azizian, Nassir Navab [arXiv]

Discovering Internal Representations from Object-CNNs Using Population Encoding
Jianyu Wang, Zhishuai Zhang, Cihang Xie, Vittal Premachandran, Alan Yuille [arXiv]

Journal Publications

Empirical Minimum Bayes Risk Prediction
Vittal Premachandran, Daniel Tarlow, Alan L. Yuille, Dhruv Batra
Transactions on Pattern Analysis and Machine Intelligence (TPAMI), 2016,

Perceptually Motivated Shape Context Which Uses Shape Interiors
Vittal Premachandran and Ramakrishna Kakarala
Pattern Recognition, 2013,
[ArXiv]

Conference Publications

• PASCAL Boundaries: A Semantic Boundary Dataset with A Deep Semantic Boundary Detector
  Vittal Premachandran, Boyan Bonev, Xiaochen Lian and Alan L. Yuille
  Winter Conference on Applications of Computer Vision (WACV), 2017, Santa Rosa, California. [PDF]
  
  
  
  
  
• Empirical Minimum Bayes Risk Prediction:
  How to extract an extra few % performance from vision models with just three more parameters,
  Vittal Premachandran, Daniel Tarlow and Dhruv Batra
  IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2014, Columbus, Ohio. [Project Page]
  
  
  
• What parts of a shape are discriminative?,[Best Student Paper Award!]
  Vittal Premachandran and Ramakrishna Kakarala
  IEEE International Conference on Image Processing (ICIP), 2013. [PDF] [Presentation Slides]
  
  
  
• Consensus of k-NNs for Robust Neighborhood Selection on Graph-Based Manifolds,
  Vittal Premachandran and Ramakrishna Kakarala
  IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2013, Portland, Oregon. [PDF]
  
  
  
  
  
• Three-dimensional bilateral symmetry plane estimation in the phase domain,
  Ramakrishna Kakarala, Prabhu Kaliamoorthi*, and Vittal Premachandran*
  IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2013, Portland, Oregon. [PDF]
  * indicates equal contribution
  
  
  
  
• Can relative skill be determined from a photographic portfolio?,
  Abhishek Agrawal, Vittal Premachandran, Rajesh Somavarapu, Ramakrishna Kakarala
  SPIE Conference on Electronic Imaging, 2013. [PDF] [SPIE]
  
  
  
  
  
• Dense sampling of shape interiors for improved representation,
  Vittal Premachandran and Ramakrishna Kakarala,
  SPIE Conference on Electronic Imaging, 2013. [PDF] [SPIE]
  
  
  
  
• Improving Shape Context Using Geodesic Information And Refection Invariance,
  Vittal Premachandran and Ramakrishna Kakarala,
  SPIE Conference on Electronic Imaging, 2013. [PDF] [SPIE]
  
  
  
  
  
• Comparing automated and human ratings of photographic aesthetics,
  Ramakrishna Kakarala, Todd Sachs, and Vittal Premachandran,
  2011 Color Imaging Conference, 7-11 Nov, San Jose CA. [PDF] [Web Link]
  
  
  
  
  
  
• "Measuring the Effectiveness of Bad Pixel Detection Algorithms Using the ROC Curve",
  Vittal Premachandran and Ramakrishna Kakarala,
  IEEE Transactions on Consumer Electronics, November 2010. [PDF] [IEEE Link]