YOLO Meets Mixture-of-Experts: Adaptive Expert Routing for Robust Object Detection

TL;DR

This paper introduces a Mixture-of-Experts framework with adaptive routing for YOLOv9-T, enhancing object detection accuracy and robustness by enabling dynamic feature specialization among multiple experts.

Contribution

It proposes a novel adaptive expert routing mechanism within a Mixture-of-Experts framework for YOLO-based object detection, improving performance over single-model approaches.

Findings

Higher mean Average Precision (mAP) and Average Recall (AR) compared to baseline YOLOv9-T

Demonstrates robustness and improved detection accuracy through expert specialization

Effective dynamic routing among experts enhances overall detection performance

Abstract

This paper presents a novel Mixture-of-Experts framework for object detection, incorporating adaptive routing among multiple YOLOv9-T experts to enable dynamic feature specialization and achieve higher mean Average Precision (mAP) and Average Recall (AR) compared to a single YOLOv9-T model.