search

Holliemin9090 / Mammographic-mass-CAD-via-pseudo-color-mammogram-and-Mask-R-CNN

Fully automatic computer-aided mass detection and segmentation via pseudo-color mammograms and Mask R-CNN
http://arxiv.org/abs/1906.12118
Other
24 stars 9 forks source link

Are you using all the 410 images of the dataset? #6

Closed latoa closed 4 years ago

Holliemin9090 commented 4 years ago

Yes. The whole dataset was split into 60% for training, 20% for validation and 20% for testing.

latoa commented 4 years ago

I have this results after the first trainning: Sensitivity:0.63636 False positive per image:0.32308 Dice:0.88455@0.0592 Any comment about? Sensitivity to low perhaps!?

Holliemin9090 commented 4 years ago

There is a parameter DETECTION_MIN_CONFIDENCE in the mamo_CAD.py file which is the confidence threshold for detection. If you set this parameter lower, the sensitivity would normally increase but also with a higher false positive rate.