文章目錄
- 題目：PFA-ScanNet: Pyramidal Feature Aggregation with Synergistic Learning for Breast Cancer Metastasis Analysis
- 時間：2019
- 會議：MICCAI
- 研究機構：港中文/中科院
1 縮寫 & 引用
- TNM: tumor, node, and distant metastasis (TNM)
- PFA: Pyramidal Feature Aggregation
- ROI: Region of Interest
2 abstract & introduction
本篇論文的主要貢獻：
- 金字塔特徵聚合Pyramidal Feature Aggregation
- 具有語義指導的協同學習
synergistic learning for training the main detector and extra decoder with semantic guidance
- 池化層提高速度
high-efficiency inference mechanism is designed with dense pooling layers
本篇論文的目的：乳腺癌轉移分析檢測，檢測轉移的存在，並測量到四個轉移類別的程度
主要難點
- 圖像像素多
- 正常和癌變區域之間存在hard mimic
- 不同轉移類型之間的顯著大小差異
2.1 相關工作
- 教師學習網絡、遷移學習
Invasive cancer detection utilizing compressed convolutional neural network and transfer learning 2018 MICCAI - 全卷積網絡
Fast scannet: fast and dense analysis of multi-gigapixel whole-slide images for cancer metastasis detection 2019 IEEE Trans. Med. Imaging 1 - 條件隨機場考慮空間相關性
Cancer metastasis detection with neural conditional random field 2018 arXiv
3 方法
3.1 Pyramidal Feature Aggregation金字塔特徵聚合
- 通過大的卷積核獲得大的感受域
- 15x15的卷積核拆成1x15+15x1和15x1+1x15兩個卷積
To further reduce the computation burden and number of parameters
3.2 FCN與池化層
The pooling strides {128,64,32} are associated with feature level {3,4,5} in the training phase and will be converted to {128/α,64/α,32/α} in the inference phase
It allows dense and fast predictions when α increases
- L_p: 訓練patch的size
- L_R: 推理input size
- S_p: scanning stide
- S_R: scanning stide for refetching ROIs
- L_m: size of predicted probability tile
Our model falls into the category of FCN architecture, which is equivalent to a patch-based CNN with input size Lp and scanning stride Sp , but the inference speed becomes much faster by removing redundant computations of overlaps
3.3 語義指導與協同學習
- BM用殘餘結構對邊界對準進行建模
BM models the boundary alignment in a residual structure to take advantage of the local contextual information and localization cue
- 同時對detector和decoder訓練
- 截斷形式的交叉熵同時減少segmentation loss和分類loss
it is hard to minimize the classification loss and segmentation loss simultaneously in one iteration
The segmentation loss will clip outliers at the truncated point γ ∈ [0,0.5] when p(t|x;W) < γ
p(t|x;W) is the predicted probability for the ground truth label t given the input pixel x
When γ=0 it will degrade into binary cross-entropy
4 實驗結果
- 數據集：Camelyon16和Camelyon17
- 隨機森林分4類：正常、ITC、Micro、Macro
- 精度高，速度快
