The burgeoning field of camouflaged object detection(COD)seeks to identify objects that blend into their *** the impressive performance of recent learning-based models,their robustness is limited,as existing methods m...
详细信息
The burgeoning field of camouflaged object detection(COD)seeks to identify objects that blend into their *** the impressive performance of recent learning-based models,their robustness is limited,as existing methods may misclassify salient objects as camouflaged ones,despite these contradictory *** limitation may stem from the lack of multipattern training images,leading to reduced robustness against salient *** overcome the scarcity of multi-pattern training images,we introduce CamDiff,a novel approach inspired by AI-Generated Content(AIGC).Specifically,we leverage a latent diffusion model to synthesize salient objects in camouflaged scenes,while using the zero-shot image classification ability of the Contrastive Language-Image Pre-training(CLIP)model to prevent synthesis failures and ensure that the synthesized objects align with the input ***,the synthesized image retains its original camouflage label while incorporating salient objects,yielding camouflaged scenes with richer *** results of user studies show that the salient objects in our synthesized scenes attract the user’s attention more;thus,such samples pose a greater challenge to the existing COD *** CamDiff enables flexible editing and effcient large-scale dataset generation at a low *** significantly enhances the training and testing phases of COD baselines,granting them robustness across diverse *** newly generated datasets and source code are available at https://***/drlxj/CamDiff.
objectdetection models typically accomplish two tasks, the localization and classification of objects. object localization algorithms have come a long way in the advancement of computer vision applications. However, ...
详细信息
ISBN:
(纸本)9781665408981
objectdetection models typically accomplish two tasks, the localization and classification of objects. object localization algorithms have come a long way in the advancement of computer vision applications. However, objectdetection models still face challenges in localizing objects that are concealed in a scene. In this paper, we investigate the localization block of different object detectors when detecting these types of objects. Our results show that for localizing objects with high color and texture similarity with its immediate background, region proposal networks perform better than single-stage networks. lb relate such blending characteristics of objects with its environments, as one possible cause of misdetection, we examine the characteristics of commonly misdetected objects with high color and texture similarity with its background. Statistical study reveals that commonly misdetected objects have significantly similar color and texture features than those of the detected objects. This implies that color and texture similarity relate to misdetection in objectdetection. To the best of our knowledge, this work presents the first empirical study on localizing objects with high color and similarity with its background.
Camouflage objectdetection (COD) aims to detect camouflagedobjects hidden in the background region in an image. The difficulty of COD lies in the fact that camouflagedobjects are often accompanied with weak bound-a...
详细信息
Camouflage objectdetection (COD) aims to detect camouflagedobjects hidden in the background region in an image. The difficulty of COD lies in the fact that camouflagedobjects are often accompanied with weak bound-aries, low contrast, and similar patterns to the background. Although various methods have been proposed to ad -dress these challenges, they still suffer from coarse object boundaries. In this work, we design a novel boundary guidance network for COD, which follows a two-step framework: localization and refinement. Firstly, an Initial Localization Decoder is proposed to capture multi-scale cues by embedding a Hierarchical-Split Convolution block. After obtained the coarse localization of the camouflagedobject, we further propose a Residual Refinement Decoder to fix the missing object parts and boundary details progressively. Each of the proposed decoder consists of a region branch and a boundary branch for objectdetection and boundary detection respectively. To suffi-ciently leverage their complementary features, we design a novel Boundary-Guide-Region module. Benefiting from the guidance of the boundary feature, the region branch can focus on the inside parts of the boundary for residual learning, thus leads to more accurate detection. Extensive experimental results on four benchmark datasets demonstrate that our method outperforms existing state-of-the-art algorithms in both object accuracy and boundary accuracy with real-time speed. (c) 2021 Elsevier B.V. All rights reserved.
暂无评论