September 20, 2018: Seminar by Pulak Purkait

Synthetic View Generation for Absolute Pose Regression and Image Synthesis
by
Pulak Purkait
Toshiba Research Europe Ltd.
Cambridge, UK
Email: pulak.cv@gmail.com
Website: http://www.pulakpurkait.com/

 
Date   : September 20, 2018
Time   : 12:00 Noon
Venue  : ECSU Seminar Room, 9th Floor, S. N. Bose Bhavan
 
Abstract
 
Image based localization is one of the important problems in computer vision due to its wide applicability in robotics, augmented reality, and autonomous systems. There is a rich set of methods described in the literature on how to geometrically register a 2D image w.r.t. a 3D model. In particular, data augmentation methods such as synthetic image generation have been shown to be useful for this task. In this work, we propose a synthetic data augmentation technique and design a deep neural network, that can be trained to estimate the absolute pose of an image from synthesized sparse feature descriptors. Our choice of using sparse feature descriptors has two major advantages: first, our network is significantly smaller than the CNNs proposed in the literature for this task—thereby making our approach more efficient and scalable. Second—and more importantly—, usage of sparse features allows to augment the training data with synthetic viewpoints, which leads to substantial improvements in the generalization performance to unseen poses. The synthetic views are further employed to augment realistic RGB images which again surpasses recent deep learning based synthetic image generation technique. A detailed analysis of the proposed networks and a rigorous evaluation on the existing datasets are provided to support our method.
 
 
All are cordially invited.
 
 
Dipti Prasad Mukherjee
Head
Electronics and Communication Sciences Unit