• Home
  • Search
  • Enhanced Feature Pyramid Networks by Feature Aggregation Module and Refinement Module
  • Cite Icon7
  • https://doi.org/10.1109/hsi49210.2020.9142674Copy DOI Icon

Enhanced Feature Pyramid Networks by Feature Aggregation Module and Refinement Module

  • Jun 1, 2020
  • Xuan-Thuy Vo +1 more
Show More
  • Abstract
  • Literature Map
  • References
  • Citations
  • Similar Papers
Abstract

Feature pyramids executing refinements on the raw feature maps produced by the backbone (e.g., ResNet, VGG) are universally employed in object detection tasks (e.g., Faster R-CNN, Mask R-CNN, YOLO, SSD, RetinaNet) to mitigate scale variation problem. Although these object detections with feature pyramids accomplish a boost in accuracy without compromising speed, they have some limitations since that they only naturally design the feature pyramid with consecutive scales, the pyramidal architecture of the backbone, which are initially constructed for the classification task. This problem leads to the feature imbalance between high-level features and low-level features in object detection. In this work, the proposed method introduces Feature Aggregation Module (FAM) and Refinement Module (RM) to obtain more powerful feature pyramids for predicting objects of different scales. First, the multi-level feature maps (i.e., multiple layers) extracted by the backbone network are aggregated as the basic feature. Second, the basic feature is enhanced by a refinement module exploiting long-range dependency. Three, to construct a feature pyramid for object detection, the proposed FAM is used by converting the basic feature (after utilizing a refinement module) into multi-level features. Finally, refined multi-level features and raw features generated by the backbone could be enhanced through shortcut connections to capture more representative. To perform the efficiency, the proposed method integrates the FAM and the RAM into the architecture of Faster R-CNN called EFPN Faster R-CNN. Especially on the MS-COCO dataset, EFPN Faster R-CNN achieves 2.2 points higher Average Precision (AP) than FPN Faster R-CNN without bells and whistles.

Similar Papers
  • Research Article
  • Citations935

M2Det: A Single-Shot Object Detector Based on Multi-Level Feature Pyramid Network

  • Jul 17, 2019
  • Proceedings of the AAAI Conference on Artificial Intelligence
  • Qijie Zhao +6
  • Conference Article
  • Citations1

Learning Richer Features in Deep CNN for Object Detection

  • Oct 01, 2020
  • Yi Li +3
  • Research Article
  • Citations12

Aggregated Residual Dilation-Based Feature Pyramid Network for Object Detection

  • Jan 01, 2019
  • IEEE Access
  • Xiaotong Zhao +5
  • Dissertation

Deep learning for small object detection in images

  • Jul 01, 2020
  • Yang Liu
  • PDF
  • Research Article
  • Citations36

Fusing attention mechanism with Mask R-CNN for instance segmentation of grape cluster in the field.

  • Jul 22, 2022
  • Frontiers in plant science
  • Lei Shen +6
  • Research Article
  • Citations10

A Multi-strategy Region Proposal Network

  • Jun 28, 2018
  • Expert Systems with Applications
  • Yu-Peng Chen +3
  • Research Article

Privacy‐protected object detection through trustworthy image fusion

  • May 09, 2024
  • International Journal of Network Management
  • Chao Zhang +4
  • Research Article
  • Citations38

Guiding Clean Features for Object Detection in Remote Sensing Images

  • Jan 01, 2022
  • IEEE Geoscience and Remote Sensing Letters
  • Gong Cheng +5
  • Research Article
  • Citations10

Complementary Feature Pyramid Network for Object Detection

  • May 31, 2023
  • ACM Transactions on Multimedia Computing, Communications, and Applications
  • Jin Xie +5
  • Research Article
  • Citations14

Ganster R-CNN: Occluded Object Detection Network Based on Generative Adversarial Nets and Faster R-CNN

  • Jan 01, 2022
  • IEEE Access
  • Kelei Sun +2
  • Research Article
  • Citations6

A Flying Bird Object Detection Method for Surveillance Video

  • Jan 01, 2024
  • IEEE Transactions on Instrumentation and Measurement
  • Zi-Wei Sun +3
  • Conference Article
  • Citations6

Clustering Anchor for Faster R-CNN to Improve Detection Results

  • Jun 01, 2020
  • Han Weiyue +1
  • Research Article
  • Citations4

Parallel channel and position attention-guided feature pyramid for pig face posture detection

  • Jan 01, 2022
  • International Journal of Agricultural and Biological Engineering
  • Zhiwei Hu +2
  • Research Article
  • Citations1

A DROPOUT TECHNIQUE STUDY FOR THE FASTER R-CNN DETECTORS WITH PRETRAINED CONVOLUTIONAL NEURAL NETWORKS FOR DETECTING VERY SIMPLE OBJECTS THAT CAN BE MASKED

  • Jan 01, 2018
  • Application Mathematics and Informatics
  • V Romanuke
  • PDF
  • Research Article
  • Citations21

3D IFPN: Improved Feature Pyramid Network for Automatic Segmentation of Gastric Tumor.

  • May 20, 2021
  • Frontiers in Oncology
  • Haimei Li +12
Cactus Communications logo

Copyright 2026 Cactus Communications. All rights reserved.