| DC 欄位 |
值 |
語言 |
| DC.contributor | 資訊工程學系 | zh_TW |
| DC.creator | 林宛儀 | zh_TW |
| DC.creator | Wan Yi Lin | en_US |
| dc.date.accessioned | 2022-7-21T07:39:07Z | |
| dc.date.available | 2022-7-21T07:39:07Z | |
| dc.date.issued | 2022 | |
| dc.identifier.uri | http://ir.lib.ncu.edu.tw:88/thesis/view_etd.asp?URN=109522127 | |
| dc.contributor.department | 資訊工程學系 | zh_TW |
| DC.description | 國立中央大學 | zh_TW |
| DC.description | National Central University | en_US |
| dc.description.abstract | 本篇論文針對圖像檢索(Image retrieval )的任務上提出了一個新的損失函數。此方法基於Proxy_NCA以及Proxy_Anchor的方法上加上了多個代表點的方法，來提升樣本的豐富性。使得Batch size減少的情況下也能達到跟原來較大的batch size一樣的效果。並且使用SoftMax函數對類內代表點做加權。使得重要的代表點能得到更多的學習資源。除此損失函數地改良之外，也對現有的ResNet50進行了修改，只使用RestNet50的前三層做為特徵擷取，取消了ResNet50第三層的下採樣。並且加入了Attention機制取代原本ResNet50的第四層。Attention使用了SoftPlus函數對特徵圖的特徵做加權。使得重要的特徵能更明顯，不重要的特徵減少關注度。 相較於傳統Attention使用SoftMax函數能得到更好的效果。不管是新提出的損失函數，或是改良過後的ResNet50都相較於原始方法Recall@1都有很大的提升。 | zh_TW |
| dc.description.abstract | In this paper, we propose a new loss function for Image Retrieval task. The new loss function makes an improvement based on Proxy-NCA and Proxy-Anchor Loss by adopting multiple proxies, to promote positive sample variety. Its shows better performance than Proxy-Anchor Loss even in the small batch size. Besides, we weighted intra-class proxy by SoftMax function to make important samples receive a higher gradient while training. In addition, we make some changes on ResNet50 by only using the first three-layer and adding a new attention module by using SoftPlus function to replace SoftMax. Finally, we obtain well results on recall@1 via our new method. | en_US |
| DC.subject | 度量學習 | zh_TW |
| DC.subject | 距離學習 | zh_TW |
| DC.subject | 圖像檢索 | zh_TW |
| DC.subject | 細粒度圖像 | zh_TW |
| DC.subject | 卷積神經網路 | zh_TW |
| DC.subject | Deep Metric Learning | en_US |
| DC.subject | Distance metric learning | en_US |
| DC.subject | Image Retrieval | en_US |
| DC.subject | Fine-grained | en_US |
| DC.subject | Convention Network | en_US |
| DC.title | Multi-Proxy Loss:基於度量學習提出之損失函數用於細粒度圖像檢索 | zh_TW |
| dc.language.iso | zh-TW | zh-TW |
| DC.title | Multi-Proxy Loss: For Deep Metric Learning on Fine-grained Image Retrieval | en_US |
| DC.type | 博碩士論文 | zh_TW |
| DC.type | thesis | en_US |
| DC.publisher | National Central University | en_US |