Learning deep representations of fine-grained visual descriptions
Abstract 最先進的zero-shot視覺識別將學習視爲圖像和補充信息的聯合問題。其中對視覺特徵來說最有效的補充信息是屬性-描述類與類之間的共享特徵的手動編碼向量。儘管算法表現很好，但是屬性任然是有侷限的： 更細粒度的識別需要相當多的屬性 屬性不提供自然語言界面（attributes do not provide a natural language interface）（不能顯式的表示？）
相關文章
- 1. Learning like a Child: Fast Novel Visual Concept Learning from Sentence Descriptions of Images
- 2. End-to-End Learning of Deep Visual Representations for Image Retrieval
- 3. Inspirational Applications of Deep Learning
- 4. [解讀] Deep Clustering for Unsupervised Learning of Visual Features
- 5. A Simple Framework for Contrastive Learning of Visual Representations
- 6. Applied Deep Learning Resources
- 7. Deep Clustering for Unspervised Learning of Visual Features
- 8. Learning Distributed Representations of Sentences from...
- 9. DeepWalk: Online Learning of Social Representations-2
- 10. 【論文筆記】Joint Unsupervised Learning of Deep Representations and Image Clusters
- 更多相關文章...
- • XSLT
- • XSLT
- • Java Agent入門實戰（一）-Instrumentation介紹與使用
- • Java Agent入門實戰（三）-JVM Attach原理與使用
相關標籤/搜索
每日一句
-
每一个你不满意的现在，都有一个你没有努力的曾经。
歡迎關注本站公眾號,獲取更多信息
相關文章
- 1. Learning like a Child: Fast Novel Visual Concept Learning from Sentence Descriptions of Images
- 2. End-to-End Learning of Deep Visual Representations for Image Retrieval
- 3. Inspirational Applications of Deep Learning
- 4. [解讀] Deep Clustering for Unsupervised Learning of Visual Features
- 5. A Simple Framework for Contrastive Learning of Visual Representations
- 6. Applied Deep Learning Resources
- 7. Deep Clustering for Unspervised Learning of Visual Features
- 8. Learning Distributed Representations of Sentences from...
- 9. DeepWalk: Online Learning of Social Representations-2
- 10. 【論文筆記】Joint Unsupervised Learning of Deep Representations and Image Clusters