EXTD: Extremely Tiny Face Detector via Iterative Filter Reuse

TL;DR

EXTD introduces an extremely lightweight multi-scale face detector that reuses a shared shallow backbone iteratively, achieving comparable accuracy to larger models with significantly fewer parameters.

Contribution

The paper presents a novel iterative backbone sharing approach for multi-scale face detection, drastically reducing model size while maintaining high performance.

Findings

Achieves comparable face detection performance to larger models.

Uses less than 0.1 million parameters.

Handles faces across various scales and conditions.

Abstract

In this paper, we propose a new multi-scale face detector having an extremely tiny number of parameters (EXTD),less than 0.1 million, as well as achieving comparable performance to deep heavy detectors. While existing multi-scale face detectors extract feature maps with different scales from a single backbone network, our method generates the feature maps by iteratively reusing a shared lightweight and shallow backbone network. This iterative sharing of the backbone network significantly reduces the number of parameters, and also provides the abstract image semantics captured from the higher stage of the network layers to the lower-level feature map. The proposed idea is employed by various model architectures and evaluated by extensive experiments. From the experiments from WIDER FACE dataset, we show that the proposed face detector can handle faces with various scale and conditions, and achieved comparable performance to the more massive face detectors that few hundreds and tens times heavier in model size and floating point operations.