[ICCV 2019] Weakly Supervised Object Detection With Segmentation Collaboration
/Li_Weakly_Supervised_Object_Detection_With_Segmentati ...
原文：[paper reading] C-MIL: Continuation Multiple Instance Learning for Weakly Supervised Object Detection CVPR2019
MIL陷入局部最優，檢測到局部，無法完整的檢測到物體。將instance划分為空間相關和類別相關的子集。在這些子集中定義一系列平滑的損失近似代替原損失函數，優化這些平滑損失。 C MIL learns instance subsets, where the instances are spatially related, i.e., overlapping with each other, and ...
2019-04-14 23:52 0 643 推薦指數：
相關推薦
論文閱讀筆記五十三：Libra R-CNN: Towards Balanced Learning for Object Detection(CVPR2019)
論文原址：https://arxiv.org/pdf/1904.02701.pdf github:https://github.com/OceanPang/Libra_R-CNN 摘要 ...
[CVPR2015] Is object localization for free? – Weakly-supervised learning with convolutional neural networks論文筆記
Is object localization for free? –Weakly-supervised learning with convolutional neural networks. Maxime Oquab, Leon Bottou, Ivan Laptev, Josef ...
[CVPR2017] Deep Self-Taught Learning for Weakly Supervised Object Localization 論文筆記
http://openaccess.thecvf.com/content_cvpr_2017/papers/Jie_Deep_Self-Taught_Learning_CVPR_2017_paper.pdf Deep Self-Taught Learning for Weakly ...
多示例學習 multiple instance learning (MIL)
多示例學習：包(bags) 和 示例 (instance). 包是由多個示例組成的，舉個例子，在圖像分類中，一張圖片就是一個包，圖片分割出的patches就是示例。在多示例學習中，包帶有類別標簽而示例不帶類別標簽，最終的目的是給出對新的包的類別預測。 多示例學習是弱監督學習中的一個 ...
Paper Reading: Relation Networks for Object Detection
Relation Networks for Object Detection筆記 寫在前面：關於這篇論文的背景知識，請參考我前面的兩篇隨筆（《關於目標檢測》和《關於注意力機制》） 摘要： 所有最先進的物體檢測系統仍然依賴於單獨識別物體實例, 在學習過程中並沒有利用它們的關系 ...
論文閱讀筆記五十一：CenterNet: Keypoint Triplets for Object Detection(CVPR2019)
論文鏈接：https://arxiv.org/abs/1904.08189 github：https://github.com/Duankaiwen/CenterNet 摘要 ...
論文閱讀筆記四十八：Bounding Box Regression with Uncertainty for Accurate Object Detection(CVPR2019)
論文原址：https://arxiv.org/pdf/1809.08545.pdf github：https://github.com/yihui-he/KL-Loss 摘要 ...