CamoFormer: Masked Separable Attention for Camouflaged Object Detection

Yin, Bowen; Zhang, Xuying; Hou, Qibin; Bo-Yuan, Sun,; Fan, Deng-Ping; Gool, Luc Van

doi:10.48550/arxiv.2212.06570

Cited by 7 publications

(12 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…And [39] introduces an explicit coarse-to-fine detection and iterative refinement strategy to refine the segmentation effect. Compared to convolutional neural networks, the vision transformer can encode global contextual information more effectively and also change the way existing methods [13], [14], [40], [41], [42], [43] perceive contextual information. Specifically, [14], [43] are based on the strategy of dividing and merging foreground and background to accurately distinguish highly similar foreground and background.…”

Section: Camouflaged Object Detection (Cod)mentioning

confidence: 99%

“…Compared to convolutional neural networks, the vision transformer can encode global contextual information more effectively and also change the way existing methods [13], [14], [40], [41], [42], [43] perceive contextual information. Specifically, [14], [43] are based on the strategy of dividing and merging foreground and background to accurately distinguish highly similar foreground and background. [13] utilizes the multi-scale correlation aggregation and the spatiotemporal transformer to construct short-term and long-term information interactions, and optimize temporal segmentation.…”

Section: Camouflaged Object Detection (Cod)mentioning

confidence: 99%

“…NC4K is another large-scale COD testing dataset including 4,121 images from the Internet. Following the data partition of [14], [15], [27], [33], [37], we use all images with camouflaged objects in the experiments, in which 3,040 images from COD10K and 1,000 images from CAMO for training, and the rest ones for testing.…”

Section: Datasetsmentioning

confidence: 99%

“…The proposed camouflaged object detector is implemented with PyTorch [83]. And the training settings are on par with recent best practices [13], [14], [15], [27], [33], [37]. The encoder is initialized with the parameters of ResNet [66], Ef-ficientNet [67] and PVTv2 [68] pre-trained on ImageNet, and the remaining parts are randomly initialized.…”

Section: Implementation Detailsmentioning

confidence: 99%

“…However, as shown in Fig. 1, the existing best solutions for image [1], [14], [15] and video [13] COD tasks have different information emphases (i.e., the appearance information of static images and the motion information of videos), resulting in incompatibility in architecture design and feature pipeline between the two types of tasks. Taking these into mind, in this paper, we summarize the COD issue into two aspects: 1) How to accurately locate camouflaged objects under conditions of inconspicuous appearance and various scales?…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Zoom In and Out: A Mixed-scale Triplet Network for Camouflaged Object Detection

Pang

Zhao

Xiang

et al. 2022

2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

187

View full text Add to dashboard Cite

Recent camouflaged object detection (COD) attempts to segment objects visually blended into their surroundings, which is extremely complex and difficult in real-world scenarios. Apart from the high intrinsic similarity between camouflaged objects and their background, objects are usually diverse in scale, fuzzy in appearance, and even severely occluded. To this end, we propose an effective unified collaborative pyramid network which mimics human behavior when observing vague images and videos, i.e., zooming in and out. Specifically, our approach employs the zooming strategy to learn discriminative mixed-scale semantics by the multi-head scale integration and rich granularity perception units, which are designed to fully explore imperceptible clues between candidate objects and background surroundings. The former's intrinsic multi-head aggregation provides more diverse visual patterns. The latter's routing mechanism can effectively propagate inter-frame difference in spatiotemporal scenarios and adaptively ignore static representations. They provides a solid foundation for realizing a unified architecture for static and dynamic COD. Moreover, considering the uncertainty and ambiguity derived from indistinguishable textures, we construct a simple yet effective regularization, uncertainty awareness loss, to encourage predictions with higher confidence in candidate regions. Our highly task-friendly framework consistently outperforms existing state-of-the-art methods in image and video COD benchmarks. The code will be available at https://github.com/lartpang/ZoomNeXt.

show abstract

Section: Camouflaged Object Detection (Cod)mentioning

confidence: 99%

Section: Camouflaged Object Detection (Cod)mentioning

confidence: 99%

Section: Datasetsmentioning

confidence: 99%

Section: Implementation Detailsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Zoom In and Out: A Mixed-scale Triplet Network for Camouflaged Object Detection

Pang

Zhao

Xiang

et al. 2022

2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

187

View full text Add to dashboard Cite

show abstract

Multi-scale pooling learning for camouflaged instance segmentation

Li,

Jiao,

Yue

et al. 2024

Appl Intell

View full text Add to dashboard Cite

Research on the restoration of costumes in the paintings of ladies in the Ming Dynasty

Liu

Gao

2023

Machine Learning, Multi Agent and Cyber Physical Systems

View full text Add to dashboard Cite

Incorporating heterogeneous representations from different architectures has facilitated various vision tasks, e.g., some hybrid networks combine transformers and convolutions. However, complementarity between such heterogeneous architectures has not been well exploited in self-supervised learning. Thus, we propose Heterogeneous Self-Supervised Learning (HSSL), which enforces a base model to learn from an auxiliary head whose architecture is heterogeneous from the base model. In this process, HSSL endows the base model with new characteristics in a representation learning way without structural changes. To comprehensively understand the HSSL, we conduct experiments on various heterogeneous pairs containing a base model and an auxiliary head. We discover that the representation quality of the base model moves up as their architecture discrepancy grows. This observation motivates us to propose a search strategy that quickly determines the most suitable auxiliary head for a specific base model to learn and several simple but effective methods to enlarge the model discrepancy. The HSSL is compatible with various self-supervised methods, achieving superior performances on various downstream tasks, including image classification, semantic segmentation, instance segmentation, and object detection. Our source code will be made publicly available.

show abstract

CamoFormer: Masked Separable Attention for Camouflaged Object Detection

Cited by 7 publications

References 0 publications

Zoom In and Out: A Mixed-scale Triplet Network for Camouflaged Object Detection

Zoom In and Out: A Mixed-scale Triplet Network for Camouflaged Object Detection

Multi-scale pooling learning for camouflaged instance segmentation

Research on the restoration of costumes in the paintings of ladies in the Ming Dynasty

Contact Info

Product

Resources

About