Improving Out-of-Distribution Detection with Disentangled Foreground and Background Features
Saved in:
| Title: | Improving Out-of-Distribution Detection with Disentangled Foreground and Background Features |
|---|---|
| Authors: | Choubo Ding, Guansong Pang |
| Source: | Proceedings of the 32nd ACM International Conference on Multimedia. :8923-8931 |
| Publication Status: | Preprint |
| Publisher Information: | ACM, 2024. |
| Publication Year: | 2024 |
| Subject Terms: | FOS: Computer and information sciences, Out-of-Distribution detection, Artificial Intelligence and Robotics, Computer Vision and Pattern Recognition (cs.CV), Disentangled representations, Computer Science - Computer Vision and Pattern Recognition, Anomaly detection, 02 engineering and technology, Machine learning, Graphics and Human Computer Interfaces, 0202 electrical engineering, electronic engineering, information engineering, Computer vision, Image representation |
| Description: | Detecting out-of-distribution (OOD) inputs is a principal task for ensuring the safety of deploying deep-neural-network classifiers in open-set scenarios. OOD samples can be drawn from arbitrary distributions and exhibit deviations from in-distribution (ID) data in various dimensions, such as foreground features (e.g., objects in CIFAR100 images vs. those in CIFAR10 images) and background features (e.g., textural images vs. objects in CIFAR10). Existing methods can confound foreground and background features in training, failing to utilize the background features for OOD detection. This paper considers the importance of feature disentanglement in out-of-distribution detection and proposes the simultaneous exploitation of both foreground and background features to support the detection of OOD inputs in in out-of-distribution detection. To this end, we propose a novel framework that first disentangles foreground and background features from ID training samples via a dense prediction approach, and then learns a new classifier that can evaluate the OOD scores of test images from both foreground and background features. It is a generic framework that allows for a seamless combination with various existing OOD detection methods. Extensive experiments show that our approach 1) can substantially enhance the performance of four different state-of-the-art (SotA) OOD detection methods on multiple widely-used OOD datasets with diverse background features, and 2) achieves new SotA performance on these benchmarks. Accepted by ACM MM 2024, 9 pages |
| Document Type: | Article |
| File Description: | application/pdf |
| DOI: | 10.1145/3664647.3681614 |
| DOI: | 10.48550/arxiv.2303.08727 |
| Access URL: | http://arxiv.org/abs/2303.08727 |
| Rights: | CC BY arXiv Non-Exclusive Distribution CC BY NC ND |
| Accession Number: | edsair.doi.dedup.....78d76fa4e66e8cb6c2f394dbbe465743 |
| Database: | OpenAIRE |
Nájsť tento článok vo Web of Science | Abstract: | Detecting out-of-distribution (OOD) inputs is a principal task for ensuring the safety of deploying deep-neural-network classifiers in open-set scenarios. OOD samples can be drawn from arbitrary distributions and exhibit deviations from in-distribution (ID) data in various dimensions, such as foreground features (e.g., objects in CIFAR100 images vs. those in CIFAR10 images) and background features (e.g., textural images vs. objects in CIFAR10). Existing methods can confound foreground and background features in training, failing to utilize the background features for OOD detection. This paper considers the importance of feature disentanglement in out-of-distribution detection and proposes the simultaneous exploitation of both foreground and background features to support the detection of OOD inputs in in out-of-distribution detection. To this end, we propose a novel framework that first disentangles foreground and background features from ID training samples via a dense prediction approach, and then learns a new classifier that can evaluate the OOD scores of test images from both foreground and background features. It is a generic framework that allows for a seamless combination with various existing OOD detection methods. Extensive experiments show that our approach 1) can substantially enhance the performance of four different state-of-the-art (SotA) OOD detection methods on multiple widely-used OOD datasets with diverse background features, and 2) achieves new SotA performance on these benchmarks.<br />Accepted by ACM MM 2024, 9 pages |
|---|---|
| DOI: | 10.1145/3664647.3681614 |