ICCV2017: Unlabeled Samples Generated by GAN Improve the Person Re-Identification Baseline in Vitro
作者：Zhedong Zheng 摘要 提出一個simple的semi-supervised pipline. 該方法只使用訓練集而不需要額外的數據。該方法的主要挑戰： - （1）如何僅從訓練集獲取更多的訓練數據， - （2）如何使用新產生的數據。 本文提出用generative adversarial network(GAN)產生無標籤數據，提出label smoothing regulizat
相關文章
- 1. Unlabeled Samples Generated by GAN Improve the Person Re-identification Baseline in vitro 論文閱讀
- 2. Multi-pseudo Regularized Label for Generated Samples in Person Re-Identification（多僞正則化標籤(MpRL)、GAN）
- 3. In Vitro model驗證 | Harnessing single-cell genomics to improve the physiological fidelity of organoid
- 4. Generative Adversarial Network (GAN) papers (2018.2.4 更新）
- 5. NVIDIA新作解讀：用GAN生成高質量行人圖像，輔助行人重識別
- 6. NVIDIA開源DG-Net：用GAN作「淘寶式」換衣，輔助行人重識別
- 7. 【FSR】Feature Space Regularization for Person Re-Identification with One Sample
- 8. 生成對抗網絡GAN（四）基於GAN的行人重識別
- 9. 行人重識別(ReID)開源代碼
- 10. A Three Player GAN Generating Hard Samples To Improve Classication Networks
- 更多相關文章...
- • SQL IN 操作符 - SQL 教程
- • Swift for-in 循環 - Swift 教程
- • Java 8 Stream 教程
- • JDK13 GA發佈：5大特性解讀
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. Unlabeled Samples Generated by GAN Improve the Person Re-identification Baseline in vitro 論文閱讀
- 2. Multi-pseudo Regularized Label for Generated Samples in Person Re-Identification（多僞正則化標籤(MpRL)、GAN）
- 3. In Vitro model驗證 | Harnessing single-cell genomics to improve the physiological fidelity of organoid
- 4. Generative Adversarial Network (GAN) papers (2018.2.4 更新）
- 5. NVIDIA新作解讀：用GAN生成高質量行人圖像，輔助行人重識別
- 6. NVIDIA開源DG-Net：用GAN作「淘寶式」換衣，輔助行人重識別
- 7. 【FSR】Feature Space Regularization for Person Re-Identification with One Sample
- 8. 生成對抗網絡GAN（四）基於GAN的行人重識別
- 9. 行人重識別(ReID)開源代碼
- 10. A Three Player GAN Generating Hard Samples To Improve Classication Networks