Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A Survey of Camouflaged Object Detection and Beyond

Fengyang Xiao, Sujie Hu, Yuqi Shen, Chengyu Fang, Jinfa Huang, Chunming He, Longxiang Tang, Ziyun Yang, Xiu Li

TL;DR

This paper explores various COD methods across four domains, including both image-level and video-level solutions, from the perspectives of traditional and deep learning approaches, from the perspectives of traditional and deep learning approaches.

Abstract

Camouflaged Object Detection (COD) refers to the task of identifying and segmenting objects that blend seamlessly into their surroundings, posing a significant challenge for computer vision systems. In recent years, COD has garnered widespread attention due to its potential applications in surveillance, wildlife conservation, autonomous systems, and more. While several surveys on COD exist, they often have limitations in terms of the number and scope of papers covered, particularly regarding the rapid advancements made in the field since mid-2023. To address this void, we present the most comprehensive review of COD to date, encompassing both theoretical frameworks and practical contributions to the field. This paper explores various COD methods across four domains, including both image-level and video-level solutions, from the perspectives of traditional and deep learning approaches. We thoroughly investigate the correlations between COD and other camouflaged scenario methods, thereby laying the theoretical foundation for subsequent analyses. Beyond object-level detection, we also summarize extended methods for instance-level tasks, including camouflaged instance segmentation, counting, and ranking. Additionally, we provide an overview of commonly used benchmarks and evaluation metrics in COD tasks, conducting a comprehensive evaluation of deep learning-based techniques in both image and video domains, considering both qualitative and quantitative performance. Finally, we discuss the limitations of current COD models and propose 9 promising directions for future research, focusing on addressing inherent challenges and exploring novel, meaningful technologies. For those interested, a curated list of COD-related techniques, datasets, and additional resources can be found at https://github.com/ChunmingHe/awesome-concealed-object-segmentation

A Survey of Camouflaged Object Detection and Beyond

TL;DR

This paper explores various COD methods across four domains, including both image-level and video-level solutions, from the perspectives of traditional and deep learning approaches, from the perspectives of traditional and deep learning approaches.

Abstract

Camouflaged Object Detection (COD) refers to the task of identifying and segmenting objects that blend seamlessly into their surroundings, posing a significant challenge for computer vision systems. In recent years, COD has garnered widespread attention due to its potential applications in surveillance, wildlife conservation, autonomous systems, and more. While several surveys on COD exist, they often have limitations in terms of the number and scope of papers covered, particularly regarding the rapid advancements made in the field since mid-2023. To address this void, we present the most comprehensive review of COD to date, encompassing both theoretical frameworks and practical contributions to the field. This paper explores various COD methods across four domains, including both image-level and video-level solutions, from the perspectives of traditional and deep learning approaches. We thoroughly investigate the correlations between COD and other camouflaged scenario methods, thereby laying the theoretical foundation for subsequent analyses. Beyond object-level detection, we also summarize extended methods for instance-level tasks, including camouflaged instance segmentation, counting, and ranking. Additionally, we provide an overview of commonly used benchmarks and evaluation metrics in COD tasks, conducting a comprehensive evaluation of deep learning-based techniques in both image and video domains, considering both qualitative and quantitative performance. Finally, we discuss the limitations of current COD models and propose 9 promising directions for future research, focusing on addressing inherent challenges and exploring novel, meaningful technologies. For those interested, a curated list of COD-related techniques, datasets, and additional resources can be found at https://github.com/ChunmingHe/awesome-concealed-object-segmentation
Paper Structure (27 sections, 6 equations, 10 figures, 8 tables)

This paper contains 27 sections, 6 equations, 10 figures, 8 tables.

Table of Contents

  1. Introduction
  2. Image-level COD models
  3. Traditional COD methods
  4. Deep learning COD methods
  5. Multi-scale-context based strategy
  6. Mechanism-simulation based strategy
  7. Multi-source information fusion strategy
  8. Multi-task learning strategy
  9. Joint-SOD based strategy
  10. Novel-task setting strategy
  11. Video-level COD models
  12. Traditional VCOD methods
  13. Deep learning VCOD methods
  14. Two-step VCOD framework
  15. End-to-end VCOD framework
  16. ...and 12 more sections

Figures (10)

  • Figure 1: Differences in input samples for GOD, SOD, and COD tasks. The three dogs, from left to right, are from the GOD dataset COCOcocolin2014microsoft, the SOD dataset DUTSDUTSwang2017learning, and the COD dataset COD10Kfan2022concealed, respectively.
  • Figure 2: The bar chart illustrates the continuous growth of COD methods across four scenarios. Representative works from various periods are categorized and marked on the line graph, with different colors corresponding to each scenario.
  • Figure 3: Network architecture, learning paradigm, and supervision level for deep learning COD methods.
  • Figure 4: The ratio of deep learning COD papers analyzed in this survey to each strategy
  • Figure 5: Two types of framework for deep learning VCOD methods. See Section \ref{['section:video-deep']} for more details
  • ...and 5 more figures