# Binarized Convolutional Landmark Localizers for Human Pose Estimation and Face Alignment with Limited Resources

## TL;DR

This paper explores binarized CNN architectures for landmark localization tasks like human pose estimation and face alignment, proposing a novel multi-scale residual design that maintains high accuracy with limited computational resources.

## Contribution

It introduces the first study on neural network binarization for localization tasks and proposes a new hierarchical, parallel, multi-scale residual architecture that improves performance without increasing parameters.

## Key findings

- Achieved state-of-the-art results on challenging datasets.
- Demonstrated the effectiveness of binarized networks for localization.
- Provided comprehensive ablation studies on the proposed architecture.

## 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