Evaluation of YOLO Models with Sliced Inference for Small Object Detection

TL;DR

This paper benchmarks YOLOv5 and YOLOX models for small object detection using the VisDrone2019Det dataset, demonstrating that sliced inference and fine-tuning significantly improve detection accuracy, especially for YOLOv5 models.

Contribution

It introduces the use of sliced inference and fine-tuning for YOLO models to enhance small object detection performance.

Findings

Sliced inference increases AP50 scores across models.

YOLOv5 models benefit more from sliced inference than YOLOX.

Combining sliced fine-tuning and inference yields substantial accuracy improvements.

Abstract

Small object detection has major applications in the fields of UAVs, surveillance, farming and many others. In this work we investigate the performance of state of the art Yolo based object detection models for the task of small object detection as they are one of the most popular and easy to use object detection models. We evaluated YOLOv5 and YOLOX models in this study. We also investigate the effects of slicing aided inference and fine-tuning the model for slicing aided inference. We used the VisDrone2019Det dataset for training and evaluating our models. This dataset is challenging in the sense that most objects are relatively small compared to the image sizes. This work aims to benchmark the YOLOv5 and YOLOX models for small object detection. We have seen that sliced inference increases the AP50 score in all experiments, this effect was greater for the YOLOv5 models compared to the YOLOX models. The effects of sliced fine-tuning and sliced inference combined produced substantial improvement for all models. The highest AP50 score was achieved by the YOLOv5- Large model on the VisDrone2019Det test-dev subset with the score being 48.8.