DC Field | Value | Language |
---|---|---|
dc.contributor.author | Kim, Jun Hyuk | ko |
dc.contributor.author | Kim, Jong-Hwan | ko |
dc.contributor.author | Yoon, Inug | ko |
dc.contributor.author | Park, Gyeongmoon | ko |
dc.date.accessioned | 2020-12-19T02:50:26Z | - |
dc.date.available | 2020-12-19T02:50:26Z | - |
dc.date.created | 2020-11-25 | - |
dc.date.issued | 2020-09-07 | - |
dc.identifier.citation | The 31st British Machine Vision Conference | - |
dc.identifier.uri | http://hdl.handle.net/10203/278752 | - |
dc.description.abstract | A few-shot image classification problem aims to recognize previously unseen objects with a small amount of data. Many works have been offered to solve the problem, while a simple transfer learning method with the cosine similarity based cross-entropy loss is still powerful compared with other methods. To improve the performance, we propose a novel Non-Probabilistic Cosine similarity (NPC) loss for few-shot classification that can replace the cross-entropy loss with the cosine similarity. A key difference of NPC loss is that it uses values of inputs instead of their probabilities. By simply changing the loss function, our model avoids overfitting on a training set and performs well on few-shot tasks. Experimental results show that the model with NPC loss clearly outperforms those with other loss functions and also achieves excellent performance compared with state-of-the-art algorithms on Mini-Imagenet and CUB-200-2011 datasets. | - |
dc.language | English | - |
dc.publisher | The British Machine Vision Association and Society for Pattern Recognition | - |
dc.title | Non-Probabilistic Cosine Similarity Loss for Few-Shot Image Classification | - |
dc.type | Conference | - |
dc.type.rims | CONF | - |
dc.citation.publicationname | The 31st British Machine Vision Conference | - |
dc.identifier.conferencecountry | UK | - |
dc.identifier.conferencelocation | Virtual | - |
dc.contributor.localauthor | Kim, Jong-Hwan | - |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.