Visual Mesh: Real-time Object Detection Using Constant Sample Density

TL;DR

This paper introduces Visual Mesh, a geometric input transformation that enhances CNN-based object detection in resource-limited robotics by normalizing object density, resulting in significantly faster execution times with high accuracy.

Contribution

The paper presents Visual Mesh, a novel geometric transformation that improves CNN efficiency and accuracy for real-time object detection in robotics.

Findings

Execution times sixteen times faster than competitors

Achieves high accuracy with reduced computational complexity

Effective in resource-constrained robotic applications

Abstract

This paper proposes an enhancement of convolutional neural networks for object detection in resource-constrained robotics through a geometric input transformation called Visual Mesh. It uses object geometry to create a graph in vision space, reducing computational complexity by normalizing the pixel and feature density of objects. The experiments compare the Visual Mesh with several other fast convolutional neural networks. The results demonstrate execution times sixteen times quicker than the fastest competitor tested, while achieving outstanding accuracy.