Locate and segment atypical objects better by being instance-aware
Locate and segment atypical objects better by being instance-aware
Instance-Aware Observer Network for Out-of-Distribution Object Segmentation
arXiv paper abstract https://arxiv.org/abs/2207.08782v1
arXiv PDF paper https://arxiv.org/pdf/2207.08782v1.pdf
Recent work on Observer Network has shown promising results on Out-Of-Distribution (OOD) detection for semantic segmentation.
These methods have difficulty in precisely locating the point of interest in the image, i.e, the anomaly.
To address this … provide instance knowledge to the observer … extend the approach of ObsNet by harnessing an instance-wise mask prediction.
… use an additional, class agnostic, object detector to filter and aggregate observer predictions.
… predict an unique anomaly score for each instance in the image.
… show that … proposed method accurately disentangle in-distribution objects from Out-Of-Distribution objects on three datasets.
Stay up to date. Subscribe to my posts https://morrislee1234.wixsite.com/website/contact
Web site with my other posts by category https://morrislee1234.wixsite.com/website
LinkedIn https://www.linkedin.com/in/morris-lee-47877b7b
