AGSFCOS: Based on attention mechanism and Scale-Equalizing pyramid network of object detection

TL;DR

This paper introduces AGSFCOS, an improved anchor-free object detection model that leverages an attention mechanism and a Scale-Equalizing pyramid network to enhance feature learning and reduce semantic gaps, achieving higher accuracy on COCO.

Contribution

The paper proposes a novel combination of attention mechanisms and a SEPC network to improve feature extraction and pyramid balancing in anchor-free detection models.

Findings

Achieves 39.5% COCO AP with ResNet50 backbone.

Attention mechanism effectively captures contextual information.

SEPC network reduces semantic gap in feature pyramids.

Abstract

Recently, the anchor-free object detection model has shown great potential for accuracy and speed to exceed anchor-based object detection. Therefore, two issues are mainly studied in this article: (1) How to let the backbone network in the anchor-free object detection model learn feature extraction? (2) How to make better use of the feature pyramid network? In order to solve the above problems, Experiments show that our model has a certain improvement in accuracy compared with the current popular detection models on the COCO dataset, the designed attention mechanism module can capture contextual information well, improve detection accuracy, and use sepc network to help balance abstract and detailed information, and reduce the problem of semantic gap in the feature pyramid network. Whether it is anchor-based network model YOLOv3, Faster RCNN, or anchor-free network model Foveabox, FSAF, FCOS. Our optimal model can get 39.5% COCO AP under the background of ResNet50.