File Download
There are no files associated with this item.
Links for fulltext
(May Require Subscription)
- Publisher Website: 10.1109/TPAMI.2021.3091944
- Scopus: eid_2-s2.0-85111176427
- WOS: WOS:000853875300067
- Find via
Supplementary
- Citations:
- Appears in Collections:
Article: AutoNovel: Automatically Discovering and Learning Novel Visual Categories
Title | AutoNovel: Automatically Discovering and Learning Novel Visual Categories |
---|---|
Authors | |
Keywords | Annotations Benchmark testing classification clustering Data models deep transfer clustering incremental learning novel category discovery Ranking (statistics) Task analysis Transfer learning Visualization |
Issue Date | 2021 |
Citation | IEEE Transactions on Pattern Analysis and Machine Intelligence, 2021 How to Cite? |
Abstract | We tackle the problem of discovering novel classes in an image collection given labelled examples of other classes. We present a new approach called AutoNovel to address this problem by combining three ideas: (1) we suggest that the common approach of bootstrapping an image representation using the labeled data only introduces an unwanted bias, and that this can be avoided by using self-supervised learning to train the representation from scratch on the union of labelled and unlabelled data; (2) we use rank statistics to transfer the model's knowledge of the labelled classes to the problem of clustering the unlabelled images; and, (3) we train the data representation by optimizing a joint objective function on the labelled and unlabelled subsets of the data, improving both the supervised classification of the labelled data, and the clustering of the unlabelled data. Moreover, we propose a method to estimate the number of classes for the case where the number of new categories is not known a priori. We evaluate AutoNovel on standard classification benchmarks and substantially outperform current methods for novel category discovery. In addition, we also show that AutoNovel can be used for fully unsupervised image clustering, achieving promising results. |
Persistent Identifier | http://hdl.handle.net/10722/311524 |
ISSN | 2023 Impact Factor: 20.8 2023 SCImago Journal Rankings: 6.158 |
ISI Accession Number ID |
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Han, Kai | - |
dc.contributor.author | Rebuffi, Sylvestre Alvise | - |
dc.contributor.author | Ehrhardt, Sebastien | - |
dc.contributor.author | Vedaldi, Andrea | - |
dc.contributor.author | Zisserman, Andrew | - |
dc.date.accessioned | 2022-03-22T11:54:08Z | - |
dc.date.available | 2022-03-22T11:54:08Z | - |
dc.date.issued | 2021 | - |
dc.identifier.citation | IEEE Transactions on Pattern Analysis and Machine Intelligence, 2021 | - |
dc.identifier.issn | 0162-8828 | - |
dc.identifier.uri | http://hdl.handle.net/10722/311524 | - |
dc.description.abstract | We tackle the problem of discovering novel classes in an image collection given labelled examples of other classes. We present a new approach called AutoNovel to address this problem by combining three ideas: (1) we suggest that the common approach of bootstrapping an image representation using the labeled data only introduces an unwanted bias, and that this can be avoided by using self-supervised learning to train the representation from scratch on the union of labelled and unlabelled data; (2) we use rank statistics to transfer the model's knowledge of the labelled classes to the problem of clustering the unlabelled images; and, (3) we train the data representation by optimizing a joint objective function on the labelled and unlabelled subsets of the data, improving both the supervised classification of the labelled data, and the clustering of the unlabelled data. Moreover, we propose a method to estimate the number of classes for the case where the number of new categories is not known a priori. We evaluate AutoNovel on standard classification benchmarks and substantially outperform current methods for novel category discovery. In addition, we also show that AutoNovel can be used for fully unsupervised image clustering, achieving promising results. | - |
dc.language | eng | - |
dc.relation.ispartof | IEEE Transactions on Pattern Analysis and Machine Intelligence | - |
dc.subject | Annotations | - |
dc.subject | Benchmark testing | - |
dc.subject | classification | - |
dc.subject | clustering | - |
dc.subject | Data models | - |
dc.subject | deep transfer clustering | - |
dc.subject | incremental learning | - |
dc.subject | novel category discovery | - |
dc.subject | Ranking (statistics) | - |
dc.subject | Task analysis | - |
dc.subject | Transfer learning | - |
dc.subject | Visualization | - |
dc.title | AutoNovel: Automatically Discovering and Learning Novel Visual Categories | - |
dc.type | Article | - |
dc.description.nature | link_to_subscribed_fulltext | - |
dc.identifier.doi | 10.1109/TPAMI.2021.3091944 | - |
dc.identifier.scopus | eid_2-s2.0-85111176427 | - |
dc.identifier.eissn | 1939-3539 | - |
dc.identifier.isi | WOS:000853875300067 | - |