SFPN: Synthetic FPN for Object Detection

TL;DR

This paper introduces SFPN, a synthetic fusion pyramid network that enhances feature fusion in object detection backbones, improving accuracy across various models.

Contribution

It proposes a novel SFPN architecture that creates synthetic layers to better fuse multi-scale features in FPN-based detectors.

Findings

SFPN outperforms traditional FPN in AP score.

Improves feature fusion for lightweight CNN backbones.

Enhances object detection accuracy across different models.

Abstract

FPN (Feature Pyramid Network) has become a basic component of most SoTA one stage object detectors. Many previous studies have repeatedly proved that FPN can caputre better multi-scale feature maps to more precisely describe objects if they are with different sizes. However, for most backbones such VGG, ResNet, or DenseNet, the feature maps at each layer are downsized to their quarters due to the pooling operation or convolutions with stride 2. The gap of down-scaling-by-2 is large and makes its FPN not fuse the features smoothly. This paper proposes a new SFPN (Synthetic Fusion Pyramid Network) arichtecture which creates various synthetic layers between layers of the original FPN to enhance the accuracy of light-weight CNN backones to extract objects' visual features more accurately. Finally, experiments prove the SFPN architecture outperforms either the large backbone VGG16, ResNet50 or light-weight backbones such as MobilenetV2 based on AP score.