This paper presents a novel framework called Dual-Constraint Coarse-to-Fine Network (DCNet) for camouflaged object detection (COD). DCNet is designed to address the challenge of detecting objects that are highly similar to their surroundings by integrating two key constraints: region and boundary information. The framework consists of three main components: an Area-Boundary Decoder (ABD), an Area Search Module (ASM), and an Area Refinement Module (ARM). The ABD extracts initial region and boundary cues from multi-level features of the backbone network. The ASM uses these cues to adaptively search for coarse regions of objects, while the ARM refines these regions to identify fine regions guided by boundary cues. The deep supervision strategy ensures that the network can localize camouflaged objects accurately by aggregating multi-level features from top to bottom. Extensive experiments on three benchmark COD datasets show that DCNet outperforms 12 state-of-the-art COD methods in terms of various evaluation metrics. Additionally, DCNet demonstrates promising performance on two COD-related tasks: industrial defect detection and polyp segmentation. The paper also includes ablation studies and a failure case analysis to validate the effectiveness of the proposed modules and identify areas for future improvements.This paper presents a novel framework called Dual-Constraint Coarse-to-Fine Network (DCNet) for camouflaged object detection (COD). DCNet is designed to address the challenge of detecting objects that are highly similar to their surroundings by integrating two key constraints: region and boundary information. The framework consists of three main components: an Area-Boundary Decoder (ABD), an Area Search Module (ASM), and an Area Refinement Module (ARM). The ABD extracts initial region and boundary cues from multi-level features of the backbone network. The ASM uses these cues to adaptively search for coarse regions of objects, while the ARM refines these regions to identify fine regions guided by boundary cues. The deep supervision strategy ensures that the network can localize camouflaged objects accurately by aggregating multi-level features from top to bottom. Extensive experiments on three benchmark COD datasets show that DCNet outperforms 12 state-of-the-art COD methods in terms of various evaluation metrics. Additionally, DCNet demonstrates promising performance on two COD-related tasks: industrial defect detection and polyp segmentation. The paper also includes ablation studies and a failure case analysis to validate the effectiveness of the proposed modules and identify areas for future improvements.