Gold-YOLO: Efficient Object Detector via Gather-and-Distribute Mechanism

TL;DR

Gold-YOLO introduces a novel gather-and-distribute mechanism that enhances multi-scale feature fusion in real-time object detection, achieving higher accuracy and efficiency than previous models while incorporating unsupervised pretraining.

Contribution

The paper proposes the Gather-and-Distribute mechanism for improved feature fusion in YOLO models and implements MAE-style pretraining for the first time in YOLO-series.

Findings

Gold-YOLO-N achieves 39.9% AP on COCO val2017.

Gold-YOLO-N runs at 1030 FPS on T4 GPU.

Outperforms YOLOv6-3.0-N in accuracy and speed.

Abstract

In the past years, YOLO-series models have emerged as the leading approaches in the area of real-time object detection. Many studies pushed up the baseline to a higher level by modifying the architecture, augmenting data and designing new losses. However, we find previous models still suffer from information fusion problem, although Feature Pyramid Network (FPN) and Path Aggregation Network (PANet) have alleviated this. Therefore, this study provides an advanced Gatherand-Distribute mechanism (GD) mechanism, which is realized with convolution and self-attention operations. This new designed model named as Gold-YOLO, which boosts the multi-scale feature fusion capabilities and achieves an ideal balance between latency and accuracy across all model scales. Additionally, we implement MAE-style pretraining in the YOLO-series for the first time, allowing YOLOseries models could be to benefit from unsupervised pretraining. Gold-YOLO-N attains an outstanding 39.9% AP on the COCO val2017 datasets and 1030 FPS on a T4 GPU, which outperforms the previous SOTA model YOLOv6-3.0-N with similar FPS by +2.4%. The PyTorch code is available at https://github.com/huawei-noah/Efficient-Computing/tree/master/Detection/Gold-YOLO, and the MindSpore code is available at https://gitee.com/mindspore/models/tree/master/research/cv/Gold_YOLO.