Data Driven Exemplar Model Selection

We start with a large set of images and remove redundancy to select a compact set of exemplars (highlighted in red). Our selection also provides a ranking so that the resulting set can be used in budgeted detection/anytime scenarios. Our results show that this selection is doubly vital - gives computational speedup, and improves performance as compared to using the entire set. Using all 1684 Exemplar-SVMs gives 0.311 mAP. Our compaction method selects only 62 Exemplar-SVMs (without training all of them) with 0.52 mAP. Less is indeed more! (see Results)

People

Ishan Misra
Abhinav Shrivastava
Martial Hebert

Overview

We consider the problem of discovering discriminative exemplars suitable for object detection. Due to the diversity in appearance in real world objects, an object detector must capture variations in scale, viewpoint, illumination etc. The current approaches do this by using mixtures of models, where each mixture is designed to capture one (or a few) axis of variation. Current methods usually rely on heuristics to capture these variations; however, it is unclear which axes of variation exist and are relevant to a particular task. Another issue is that we require a large set of training images to capture such variations. Current methods do not scale to large training sets either because of training time complexity or test time complexity. In this work, we explore the idea of compactly capturing task-appropriate variation from the data itself. We propose a two stage data-driven process, which selects and learns a compact set of exemplar models for object detection. These selected models have an inherent ranking, which can be used for anytime/budgeted detection scenarios. Another benefit of our approach (beyond the computational speedup) is that the selected set of exemplar models performs better than the entire set.

Some Results


The ranking provided by our algorithm as compared to that of Sparse Modeling (Elhamifar et al., 2012) on the KITTI car dataset.	Our Method (LDA Pruning+Selection) against baselines.

Paper

Ishan Misra, Abhinav Shrivastava and Martial Hebert. Data-driven Exemplar Model Selection, Proceedings of IEEE Winter Conference on Applications of Computer Vision (WACV), 2014. [PDF] [BibTeX] (Best Student Paper Award)

Acknowledgement

This work was supported in part by NSF Grant IIS1065336 and the Siebel Scholarship.

Copyright notice

The documents contained in these directories are included by the contributing authors as a means to ensure timely dissemination of scholarly and technical work on a non-commercial basis. Copyright and all rights therein are maintained by the authors or by other copyright holders, notwithstanding that they have offered their works here electronically. It is understood that all persons copying this information will adhere to the terms and constraints invoked by each author's copyright. These works may not be reposted without the explicit permission of the copyright holder, except when identified by Creative Commons License 2.0, in which case the license applies to both the original and modified versions of the images.