TPRNet: camouflaged object detection via transformer-induced progressive refinement network

Zhang, Qiao; Ge, Yanliang; Zhang, Cong; Bi, Hongbo

doi:10.1007/s00371-022-02611-1

Cited by 34 publications

(2 citation statements)

References 50 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…TPRNet [37], FAPNet[38], ZoomNet [3], ERRNet [7], and FEDER [37]. The comparative data is obtained by retraining using the original authors' source code in the same training environment.…”

Section: Comparison With State-of-the-artsmentioning

confidence: 99%

Camouflage Object Detection Based on Multi-Task Learning

Lin,

Han

2024

Preprint

View full text Add to dashboard Cite

Camouflaged object detection(COD) is an emerging and challenging visual detection task that can recognize and segment objects that perfectly blend into their surroundings. This paper introduces a novel multi-task learning framework(DEENet). It primarily features a dual-branch structure: one branch is the edge enhancement module based on multi-head attention mechanism, and the other branch is the semantic feature extraction module. We effectively utilize the gradient information to integrate two kinds of features through the gradient-induced transition module. Extensive experiments on three challenging test datasets demonstrate that our DEENet outperforms existing state-of-the-art methods.

show abstract

“…TPRNet [37], FAPNet[38], ZoomNet [3], ERRNet [7], and FEDER [37]. The comparative data is obtained by retraining using the original authors' source code in the same training environment.…”

Section: Comparison With State-of-the-artsmentioning

confidence: 99%

Camouflage Object Detection Based on Multi-Task Learning

Lin,

Han

2024

Preprint

View full text Add to dashboard Cite

show abstract

“…Zhang et al [15] designed a COD method using a transformer-included progressive refinement network (TPRNet). TPRNet detects the semantic and spatial parameters and information for the detection process.…”

Section: Related Workmentioning

confidence: 99%