Medical Imaging (D3L3 2017 UPC Deep Learning for Computer Vision)
6 likes3,308 views
The document outlines the applications of deep learning in medical imaging, focusing on detection tasks like diabetic retinopathy and skin cancer, as well as segmentation challenges for brain tumors. It discusses various studies and architectures used in these classifications and segmentations, highlighting the benefits and obstacles faced in the field. The content is presented by Verónica Vilaplana from the Universitat Politècnica de Catalunya.
![[course site]
Verónica Vilaplana
veronica.vilaplana@upc.edu
Associate Professor
Universitat Politecnica de Catalunya
Technical University of Catalonia
Medical imaging
applications
#DLUPC](https://image.slidesharecdn.com/dlcv2017d3l3medicalimaging-170623173104/85/Medical-Imaging-D3L3-2017-UPC-Deep-Learning-for-Computer-Vision-1-320.jpg)
Medical Imaging (D3L3 2017 UPC Deep Learning for Computer Vision)
- 1. [course site] Verónica Vilaplana veronica.vilaplana@upc.edu Associate Professor Universitat Politecnica de Catalunya Technical University of Catalonia Medical imaging applications #DLUPC
- 3. Why deep learning for medical imaging? 3
- 4. Why deep learning for medical imaging? 4 ● ● ●
- 5. 5 Why deep learning for medical imaging?
- 6. Why deep learning for medical imaging? 6
- 7. Deep learning uses in medical imaging 7
- 11. Classification: Diabetic Retinopathy detection 11 Google: V. Gulshan et al, “Development and Validation of a Deep Learning Algorithm for Detection of Diabetic Retinopathy in Retinal Fundus Photographs”, Journal of American Medical Association, Dec. 2016
- 12. Classification: Diabetic Retinopathy detection 12 Szegedy et al, “Rethinking the Inception Architecture for Computer Vision”, Dec 2015
- 13. Classification: Diabetic Retinopathy detection 13
- 14. Classification: skin cancer detection 14
- 15. Classification: skin cancer detection 15
- 16. Classification: skin cancer detection 16
- 17. Classification: skin cancer detection 17 Inference classes Training classes
- 18. Classification: skin cancer detection 18
- 19. Classification: skin cancer detection 19 t-SNE visualization of the last hidden layer
- 20. Segmentation: cell segmentation in microscopy images ● ● 20Ronneberger et al, “U-Net: Convolutional Networks for Biomedical Image Segmentation”, arXiv 2015
- 21. U-Net ● ● ● 21
- 22. Segmentation and Classification with partial annotations 22
- 23. Segmentation and Classification with partial annotations 23
- 24. Segmentation and Classification with partial annotations 24* Simonyan et al, Deep inside convolutional networks: visualising image classification models and saliency maps, 2013
- 25. Segmentation: brain tumor segmentation 25
- 26. Segmentation: brain tumor segmentation 26
- 27. Segmentation: brain tumor segmentation 27 Casamitjana et al, 3D Convolutional Networks for Brain Tumor Segmentation: a comparison of multiresolution architectures, 2016. (1) Havaei et al, Brain tumor segmentation with Deep Neural Networks, 2016 (2) Kamnitsas, Efficient Multi Scale 3D CNN with fully connected CRF for Accurate Brain Lesion Segmentation, 2016
- 28. Segmentation: brain tumor segmentation 28 3D-Net1 3D-Net3 3D-Net2
- 29. Segmentation: brain tumor segmentation 29 The importance of skip connections
- 31. Questions? 31