Natural Image
Natural images, encompassing photographs and other visual data from the real world, are a central focus in computer vision research, aiming to enable machines to understand and interact with visual information as humans do. Current research emphasizes developing robust models, often leveraging architectures like Vision Transformers and diffusion models, to address challenges such as object detection, segmentation, and scene understanding in complex, diverse imagery. This work is crucial for advancing applications ranging from medical image analysis and autonomous navigation to improved image generation and quality assessment, ultimately bridging the gap between human and machine perception.
Papers
Blended Diffusion for Text-driven Editing of Natural Images
Omri Avrahami, Dani Lischinski, Ohad Fried
Instance-wise Occlusion and Depth Orders in Natural Scenes
Hyunmin Lee, Jaesik Park
OOD-CV: A Benchmark for Robustness to Out-of-Distribution Shifts of Individual Nuisances in Natural Images
Bingchen Zhao, Shaozuo Yu, Wufei Ma, Mingxin Yu, Shenxiao Mei, Angtian Wang, Ju He, Alan Yuille, Adam Kortylewski