Adaptation of MobileNetV2 for Face Detection on Ultra-Low Power Platform

TL;DR

This paper adapts MobileNetV2 architecture for face detection on ultra-low power embedded hardware, focusing on topology modifications and quantization to optimize performance within hardware constraints.

Contribution

It introduces specific adaptations to MobileNetV2 and evaluates their effects on face detection performance on low-power embedded devices.

Findings

Topology modifications improve deployment efficiency

Post-training quantization reduces model size and inference time

Adapted model achieves effective face detection on embedded hardware

Abstract

Designing Deep Neural Networks (DNNs) running on edge hardware remains a challenge. Standard designs have been adopted by the community to facilitate the deployment of Neural Network models. However, not much emphasis is put on adapting the network topology to fit hardware constraints. In this paper, we adapt one of the most widely used architectures for mobile hardware platforms, MobileNetV2, and study the impact of changing its topology and applying post-training quantization. We discuss the impact of the adaptations and the deployment of the model on an embedded hardware platform for face detection.