Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Green Video Camouflaged Object Detection

Xinyu Wang, Hong-Shuo Chen, Zhiruo Zhou, Suya You, Azad M. Madni, C. -C. Jay Kuo

TL;DR

This work proposes a green VCOD method named GreenVCOD, built upon a green ICOD method, that uses long- and short-term temporal neighborhoods (TN) to capture joint spatial/temporal context information for decision refinement.

Abstract

Camouflaged object detection (COD) aims to distinguish hidden objects embedded in an environment highly similar to the object. Conventional video-based COD (VCOD) methods explicitly extract motion cues or employ complex deep learning networks to handle the temporal information, which is limited by high complexity and unstable performance. In this work, we propose a green VCOD method named GreenVCOD. Built upon a green ICOD method, GreenVCOD uses long- and short-term temporal neighborhoods (TN) to capture joint spatial/temporal context information for decision refinement. Experimental results show that GreenVCOD offers competitive performance compared to state-of-the-art VCOD benchmarks.

Green Video Camouflaged Object Detection

TL;DR

This work proposes a green VCOD method named GreenVCOD, built upon a green ICOD method, that uses long- and short-term temporal neighborhoods (TN) to capture joint spatial/temporal context information for decision refinement.

Abstract

Camouflaged object detection (COD) aims to distinguish hidden objects embedded in an environment highly similar to the object. Conventional video-based COD (VCOD) methods explicitly extract motion cues or employ complex deep learning networks to handle the temporal information, which is limited by high complexity and unstable performance. In this work, we propose a green VCOD method named GreenVCOD. Built upon a green ICOD method, GreenVCOD uses long- and short-term temporal neighborhoods (TN) to capture joint spatial/temporal context information for decision refinement. Experimental results show that GreenVCOD offers competitive performance compared to state-of-the-art VCOD benchmarks.
Paper Structure (20 sections, 6 figures, 3 tables)

This paper contains 20 sections, 6 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Review of Related Work
  3. Image Camouflaged Object Detection
  4. Video Camouflaged Object Detection
  5. GreenVCOD Method
  6. Initial-stage Prediction Module
  7. Feature Extraction Module
  8. Multi-resolution Decision Module
  9. Temporal Refinement Module
  10. Temporal Neighborhood (TN) Prediction Cube Reconstruction
  11. Long-Term and Short-Term Decision Refinement
  12. Decision Ensemble Module
  13. Experiments
  14. Datasets
  15. Experimental Setup
  16. ...and 5 more sections

Figures (6)

  • Figure 1: The system diagram of the proposed green video camouflaged object detection method (GreenVCOD).
  • Figure 2: Comparison between traditional explicit motion detection and our proposed fixed focus temporal information extraction.
  • Figure 3: TN Prediction Cube Reconstruction: The brown frame represents the current frame, while the yellow frames represent the selected adjacent frames.
  • Figure 4: Temporal Refinement Module Diagram.
  • Figure 5: Visual comparisons on example video frames with state-of-the-art camouflage detection methods.
  • ...and 1 more figures