From 61e66ba0553a648e0959d0a65ea22edb1702756e Mon Sep 17 00:00:00 2001 From: Tanqiu Liu Date: Mon, 18 Sep 2017 22:20:33 -0500 Subject: [PATCH] Update README.md --- README.md | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/README.md b/README.md index 87c0e5d..75e411f 100644 --- a/README.md +++ b/README.md @@ -17,11 +17,11 @@ The detection process consists of 2 steps: 1) locating cell centers with a convo A convolutional neural network model was developed to automatically identify the location of the cell centers. The pipeline employed a sliding window approach for detection. That is , a small window slides across the entire image and for position output whether the position is the center of a cell or not. The training data was extracted and transformed from raw .csg files which are manually labeled by former researchers. Positive examples are obtained by generating a window centered on the center of the cell. Negative examples are obtained by random sampling outside the cells. -training examples +training examples The convolutional neural network consists of three convolutional layers and two dense layers. The CNN predicts whether the current window is a cell. -CNN structure +CNN structure For each image, the sliding window method predicts a set of points which are candidates of centers of cells. Following postprocessing of these candidates includes discard points outside the cells and merge points in the cells. Finally we get precise and robust predictions of centers of cells. @@ -40,4 +40,4 @@ The principal algorithm is : Sample results: -       +