Binarized convolutional landmark localizers for human pose estimation and face alignment with limited resources
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...
| Main Authors: | , |
|---|---|
| Format: | Conference or Workshop Item |
| Published: |
2017
|
| Online Access: | https://eprints.nottingham.ac.uk/44753/ |
| _version_ | 1848796990499979264 |
|---|---|
| author | Bulat, Adrian Tzimiropoulos, Georgios |
| author_facet | Bulat, Adrian Tzimiropoulos, Georgios |
| author_sort | Bulat, Adrian |
| building | Nottingham Research Data Repository |
| collection | Online Access |
| description | 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 |
| first_indexed | 2025-11-14T19:56:46Z |
| format | Conference or Workshop Item |
| id | nottingham-44753 |
| institution | University of Nottingham Malaysia Campus |
| institution_category | Local University |
| last_indexed | 2025-11-14T19:56:46Z |
| publishDate | 2017 |
| recordtype | eprints |
| repository_type | Digital Repository |
| spelling | nottingham-447532020-05-04T19:14:18Z https://eprints.nottingham.ac.uk/44753/ Binarized convolutional landmark localizers for human pose estimation and face alignment with limited resources Bulat, Adrian Tzimiropoulos, Georgios 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 2017-10-26 Conference or Workshop Item PeerReviewed 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. http://ieeexplore.ieee.org/document/8237662/ |
| spellingShingle | Bulat, Adrian Tzimiropoulos, Georgios Binarized convolutional landmark localizers for human pose estimation and face alignment with limited resources |
| title | Binarized convolutional landmark localizers for human pose estimation and face alignment with limited resources |
| title_full | Binarized convolutional landmark localizers for human pose estimation and face alignment with limited resources |
| title_fullStr | Binarized convolutional landmark localizers for human pose estimation and face alignment with limited resources |
| title_full_unstemmed | Binarized convolutional landmark localizers for human pose estimation and face alignment with limited resources |
| title_short | Binarized convolutional landmark localizers for human pose estimation and face alignment with limited resources |
| title_sort | binarized convolutional landmark localizers for human pose estimation and face alignment with limited resources |
| url | https://eprints.nottingham.ac.uk/44753/ https://eprints.nottingham.ac.uk/44753/ |