Lien de la note Hackmd
X-ray breast imaging
- worldwide breast cancer account for $25.2\%$ of all female cancers and 16% of cancer deaths in adult women
- mammography uses low-energy X-ray
Il faut detecter des lesions tres petites
- Certaines lesions sont begnines
Lesions
Artificial intelligence applications
- AI tools are becoming a must-have for mammography screening systems
- They can provide significant speed increase of radiologist workflow and more importantly improve quality of their job
- Possible applications:
- Negative Triage
- AI algorithm triage of negative cases in order to allow radiologist to concentrate on more important points
- AI CAD
- Detect cancer at very early stages which can not be identified by radiologist
- Negative Triage
Project description
- Perform analysis of object detection state-of-the-art methods
- Define model architecture on the example of public dataset
- Digital Database for Screening Mammography
- Implement pipeline for model training
- Perform training and optimization of the model
Question
C’est en python ?
Oui
Si j’ai bien compris le projet est a réalisé depuis le tt début ? Il y a rien de fait ?
A deja un prototype, le deep learning se developpe tres vite et veut investiguer des nouvelles propositions