Binarized Convolutional Landmark Localizers for Human Pose Estimation and Face Alignment with Limited Resources
arXiv:1703.00862 · doi:10.1109/ICCV.2017.400
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
ICCV 2017 Oral
Cited by in corpus (9)
- How far are we from solving the 2D & 3D Face Alignment problem? (and a dataset of 230,000 3D facial landmarks)
- XNect: Real-time Multi-Person 3D Motion Capture with a Single RGB Camera
- A comprehensive review of Binary Neural Network
- Multi-task head pose estimation in-the-wild
- Structured Landmark Detection via Topology-Adapting Deep Graph Learning
- Face Alignment using a 3D Deeply-initialized Ensemble of Regression Trees
- Heatmap Regression via Randomized Rounding
- Hierarchical binary CNNs for landmark localization with limited resources
- A Review of Recent Advances of Binary Neural Networks for Edge Computing