Learning Lightweight Pedestrian Detector with Hierarchical Knowledge Distillation

TL;DR

This paper introduces a hierarchical knowledge distillation framework that enables training lightweight pedestrian detectors with significantly fewer parameters while maintaining high accuracy, suitable for real-world applications.

Contribution

It proposes a novel hierarchical distillation approach that learns multi-level features at different detector stages, improving lightweight model performance.

Findings

Student model with 6x fewer parameters achieves competitive accuracy.

Hierarchical distillation improves low-level detail and high-level abstraction learning.

Framework reduces computational cost while maintaining detection performance.

Abstract

It remains very challenging to build a pedestrian detection system for real world applications, which demand for both accuracy and speed. This work presents a novel hierarchical knowledge distillation framework to learn a lightweight pedestrian detector, which significantly reduces the computational cost and still holds the high accuracy at the same time. Following the `teacher--student' diagram that a stronger, deeper neural network can teach a lightweight network to learn better representations, we explore multiple knowledge distillation architectures and reframe this approach as a unified, hierarchical distillation framework. In particular, the proposed distillation is performed at multiple hierarchies, multiple stages in a modern detector, which empowers the student detector to learn both low-level details and high-level abstractions simultaneously. Experiment result shows that a student model trained by our framework, with 6 times compression in number of parameters, still achieves competitive performance as the teacher model on the widely used pedestrian detection benchmark.