2 Hits in 2.7 sec

Iterative Saliency Enhancement using Superpixel Similarity [article]

Leonardo de Melo Joao, Alexandre Xavier Falcao
2022 arXiv   pre-print
Saliency Object Detection (SOD) has several applications in image analysis. The methods have evolved from image-intrinsic to object-inspired (deep-learning-based) models.  ...  ISESS estimates seeds for superpixel delineation from a given saliency map and defines superpixel queries in the foreground and background.  ...  (Zhang et al., 2021) proposes the Auto-Multi-Scale Fusion Network (Auto-MSFNet), which uses Network Architecture Search (NAS) as inspiration to propose an automatic fusion framework instead of trying  ... 
arXiv:2112.00665v2 fatcat:rldyrs7lnnhe3pim5j6djls7ay

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

Pang Youwei, Zhao Xiaoqi, Xiang Tian-Zhu, Zhang Lihe, Lu Huchuan
2022 arXiv   pre-print
The recently proposed camouflaged object detection (COD) attempts to segment objects that are visually blended into their surroundings, which is extremely complex and difficult in real-world scenarios.  ...  To deal with these problems, we propose a mixed-scale triplet network, ZoomNet, which mimics the behavior of humans when observing vague images, i.e., zooming in and out.  ...  Camouflaged object detection (COD) is far more complex and challenging than traditional salient object detection or other object segmentation.  ... 
arXiv:2203.02688v1 fatcat:vci4pztehjf55lpztvfa2y2kny