Bi-box Regression for Pedestrian Detection and Occlusion Estimation
一個迴歸full body，一個迴歸visible part。 分數融合方式。關鍵就是中間那個p^1 hat。其實就是把原來的正分相加，負分相加，再做softmax。 對proposal是否是正樣本判斷的規則改進：要跟visible重合區域夠大才算。
相關文章
- 1. CUHK Occlusion Dataset (for pedestrian detection)
- 2. Linear Regression and Maximum Likelihood Estimation
- 3. Caltech Pedestrian Detection Benchmark
- 4. PointSetNet：Point-Set Anchors for Object Detection,Instance Segmentation and Pose Estimation
- 5. Robust Face Alignment Under Occlusion via Regional Predictive Power Estimation
- 6. How Far are We from Solving Pedestrian Detection?
- 7. Object Detection(目標檢測神文)(二)
- 8. Deep Learning Models and Code for Pose Estimation
- 9. Pushing the Limits of Deep CNNs for Pedestrian Detection
- 10. Learning Cross-Modal Deep Representations for Robust Pedestrian Detection
- 更多相關文章...
- • Swift for 循環 - Swift 教程
- • Scala for循環 - Scala教程
- • RxJava操作符（七）Conditional and Boolean
- • 算法總結-股票買賣
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. CUHK Occlusion Dataset (for pedestrian detection)
- 2. Linear Regression and Maximum Likelihood Estimation
- 3. Caltech Pedestrian Detection Benchmark
- 4. PointSetNet：Point-Set Anchors for Object Detection,Instance Segmentation and Pose Estimation
- 5. Robust Face Alignment Under Occlusion via Regional Predictive Power Estimation
- 6. How Far are We from Solving Pedestrian Detection?
- 7. Object Detection(目標檢測神文)(二)
- 8. Deep Learning Models and Code for Pose Estimation
- 9. Pushing the Limits of Deep CNNs for Pedestrian Detection
- 10. Learning Cross-Modal Deep Representations for Robust Pedestrian Detection