論文閱讀筆記《GANomaly: Semi-Supervised Anomaly Detection via Adversarial Training》
核心思想 本文提出一種基於半監督GAN的異常檢測算法，主要的創新點在於在自動編碼器之後，又增加了一個編碼器，構成了「編碼-解碼-編碼」的結構，通過比較第二個編碼器的輸出和第一個編碼器的輸出之間的差別，來評估輸入圖像是否屬於異常圖像。網絡的訓練過程同樣也引入了對抗訓練的思想，即使用一個區分器對解碼器輸出的重構圖像進行判斷，是來自原始圖像還是重構圖像，整個網絡的結構如下圖所示 整個網絡可以分爲
相關文章
- 1. GAN異常檢測論文筆記《GANomaly: Semi-Supervised Anomaly Detection via Adversarial Training》
- 2. 【論文】Skip-GANomaly: Skip Connected and Adversarially Trained Encoder-Decoder Anomaly Detection
- 3. [論文閱讀筆記]Learning Memory-guided Normality for Anomaly Detection
- 4. 論文閱讀：Dehaze-GLCGAN: Unpaired Single Image Dehazing Via Adversarial Training
- 5. 論文譯文《ANOMALY DETECTION WITH GENERATIVE ADVERSARIAL NETWORKS》
- 6. 論文閱讀筆記《Anomaly Detection in Nanofibrous Materials by CNN-Based Self-Similarity》
- 7. 《Long Text Generation via Adversarial Training with Leaked Information》閱讀筆記
- 8. VITAL: Visual Tracking via Adversarial Learning 閱讀筆記
- 9. 論文閱讀筆記： 2019 ICCV 《Improving Adversarial Robustness via Guided Complement Entropy》
- 10. 【論文閱讀】Illuminating Pedestrians via Simultaneous Detection & Segmentation
- 更多相關文章...
- • RSS 閱讀器 - RSS 教程
- • C# 文本文件的讀寫 - C#教程
- • Tomcat學習筆記（史上最全tomcat學習筆記）
- • 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。
最新文章
歡迎關注本站公眾號,獲取更多信息
相關文章
- 1. GAN異常檢測論文筆記《GANomaly: Semi-Supervised Anomaly Detection via Adversarial Training》
- 2. 【論文】Skip-GANomaly: Skip Connected and Adversarially Trained Encoder-Decoder Anomaly Detection
- 3. [論文閱讀筆記]Learning Memory-guided Normality for Anomaly Detection
- 4. 論文閱讀：Dehaze-GLCGAN: Unpaired Single Image Dehazing Via Adversarial Training
- 5. 論文譯文《ANOMALY DETECTION WITH GENERATIVE ADVERSARIAL NETWORKS》
- 6. 論文閱讀筆記《Anomaly Detection in Nanofibrous Materials by CNN-Based Self-Similarity》
- 7. 《Long Text Generation via Adversarial Training with Leaked Information》閱讀筆記
- 8. VITAL: Visual Tracking via Adversarial Learning 閱讀筆記
- 9. 論文閱讀筆記： 2019 ICCV 《Improving Adversarial Robustness via Guided Complement Entropy》
- 10. 【論文閱讀】Illuminating Pedestrians via Simultaneous Detection & Segmentation