Abstract:In existing state-of-the-art object detectors, feature pyramid networks (FPN) and multiscale feature fusion are still typically used. The traditional FPN fusion strategy is based on the top-down fusion of high-level semantic information. The top-down fusion method generally uses upsampling based on interpolation, which often results in jagged edges, mosaic distortion, and edge blurring. Moreover, in order to improve accuracy, the FPN-based fusion strategy must add multiple top-down components for fusion, which increases computational costs and leads to a poor balance between precision and speed. In this paper, we propose a novel fusion strategy based on a backbone network. We aim to design simple and efficient components for high-quality object detection. Our proposed strategy, bi-directional skip connection FPN (BiSCFPN), consists of three components: a bi-directional skip connection (BiSC), a selective dilated convolution module (SDCM), and sub-pixel convolution (SP). The BiSC aims to enhance semantic information between different feature layers in the backbone network and simultaneously uses the SDCM to improve the receptive fields of differently sized targets in the fusion stage. Finally, SP learns the relationship between the features of upsampling and downsampling images to effectively mitigate the problems caused by the traditional interpolation method. BiSCFPN achieves an average precision of 38.2% in tests with the Microsoft Common Objects in Context (MS COCO) test-dev dataset at a real-time speed of ~50 FPS (608×608) using an Nvidia GeForce RTX 2080 Ti graphics card and significantly improves the balance between precision and speed.

Improvements Based on Feature Fusion Single Shot Multibox Detector

FSSD: Feature Fusion Single Shot Multibox Detector

TDFSSD: Top-Down Feature Fusion Single Shot MultiBox Detector

FFR-SSD: feature fusion and reconstruction single shot detector for multi-scale object detection

Improved single shot multibox detector target detection method based on deep feature fusion

Object Detection Using Deep Learning: Single Shot Detector with a Refined Feature-fusion Structure

A feature pyramid network with adaptive fusion strategy and enhanced semantic information

A Rich Feature Fusion Single-Stage Object Detector.

Object Detection Based on the Improved Single Shot MultiBox Detector

FESSD:SSD target detection based on feature fusion and feature enhancement

Enhancement-fusion Feature Pyramid Network for Object Detection

FRFB: Integrate Receptive Field Block into Feature Fusion Net for Single Shot Multibox Detector.

Attention-based Fusion Factor in FPN for Object Detection

Parallel Residual Bi-Fusion Feature Pyramid Network for Accurate Single-Shot Object Detection

A Balanced Feature Fusion SSD for Object Detection

Bi-directional Skip Connection Feature Pyramid Network and Sub-Pixel Convolution for High-Quality Object Detection

Single-Shot Object Detection via Feature Enhancement and Channel Attention

HYPER FEATURE FUSION PYRAMID NETWORK FOR OBJECT DETECTION

Small Object Detection Algorithm Based on Feature Pyramid-Enhanced Fusion SSD

Cascaded Multi-Channel Feature Fusion for Object Detection.

Adaptive Multiscale Feature for Object Detection