Binarized convolutional landmark localizers for human pose estimation and face alignment with limited resources

Bulat, Adrian and Tzimiropoulos, Georgios (2017) Binarized convolutional landmark localizers for human pose estimation and face alignment with limited resources. In: International Conference on Computer Vision (ICCV17), 22-29 Oct 2017, Venice, Italy.

[img]
Preview
PDF - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader
Download (1MB) | Preview

Abstract

Our goal is to design architectures that retain the groundbreaking performance of CNNs for landmark localization and at the same time are lightweight, compact and suitable for applications with limited computational resources. To this end, we make the following contributions: (a) we are the first to study the effect of neural network binarization on localization tasks, namely human pose estimation and face alignment. We exhaustively evaluate various design choices, identify performance bottlenecks, and more importantly propose multiple orthogonal ways to boost performance. (b) Based on our analysis, we propose a novel hierarchical, parallel and multi-scale residual architecture that yields large performance improvement over the standard bottleneck block while having the same number of parameters, thus bridging the gap between the original network and its binarized counterpart. (c) We perform a large number of ablation studies that shed light on the properties and the performance of the proposed block. (d) We present results for experiments on the most challenging datasets for human pose estimation and face alignment, reporting in many cases state-of-the-art performance. Code can be downloaded from https://www. adrianbulat.com/binary-cnn-landmarks

Item Type: Conference or Workshop Item (Paper)
Additional Information: © 2017 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. Published in 2017 IEEE International Conference on Computer Vision ISBN 9781538610329
Schools/Departments: University of Nottingham, UK > Faculty of Science > School of Computer Science
Related URLs:
Depositing User: Tzimiropoulos, Yorgos
Date Deposited: 09 Aug 2017 07:44
Last Modified: 12 Jan 2018 00:28
URI: http://eprints.nottingham.ac.uk/id/eprint/44753

Actions (Archive Staff Only)

Edit View Edit View