Towards lightweight convolutional neural networks for object detection

TL;DR

This paper introduces a lightweight, accurate, and fast object detection model optimized for embedded systems, achieving real-time inference on CPU with minimal accuracy loss.

Contribution

It proposes a new lightweight CNN-based object detection model with larger feature maps and channel reduction techniques, suitable for embedded applications.

Findings

Achieves 93.39 AP with 1.5 GFLOPs on DETRAC dataset.

First lightweight model to run in real-time on CPU.

Maintains high accuracy with reduced computational complexity.

Abstract

We propose model with larger spatial size of feature maps and evaluate it on object detection task. With the goal to choose the best feature extraction network for our model we compare several popular lightweight networks. After that we conduct a set of experiments with channels reduction algorithms in order to accelerate execution. Our vehicle detection models are accurate, fast and therefore suit for embedded visual applications. With only 1.5 GFLOPs our best model gives 93.39 AP on validation subset of challenging DETRAC dataset. The smallest of our models is the first to achieve real-time inference speed on CPU with reasonable accuracy drop to 91.43 AP.