The advances in digitalization of tissue slides and the FDA approval of digital pathology as a primary diagnostic tool in 2017 have created a lot of attention in the digital pathology field. Tissue can be stained in multiple ways, creating a variety of pathology sub-modalities depending on the stainings used: H&E or specific antibodies for HER2, Ki67...
The overwhelming majority of AI-based efforts in digital pathology focus on one single modality (H&E), with CNNs being the favorite tool. This type of approach has important limitations due to the enormous size of digital pathology images, but also because the pixel-based reasoning is not relatable for domain experts.
The proposed project involves developing a deep learning framework multimodal digital pathology possibly extending the concept of HACT-Net and addressing the following challenges:
- Multimodal representation: what is the most efficient way of representing a series of images of multiple stains?
- Multimodal alignment: how is the local information from each of these images preserved and aligned throughout the decision process?
- Multimodal fusion: how is the information from various images fused to make the best possible decision?
- ETH Master students with a background in Computer Science, Computer Vision, Medical Image or related fields.
- Solid background in machine learning/deep learning and computer vision/medical imaging.
- Strong programming skills in python.
- Practical experience with at least one deep learning framework (Tensorflow, Pytorch) are essential.
- Prior knowledge of digital pathology is welcome, but not a prerequisite.
IBM is committed to diversity at the workplace. With us you will find an open, multicultural environment. Excellent flexible working arrangements enable all genders to strike the desired balance between their professional development and their personal lives.
How to apply
If you are interested in this position, please submit your application including a recent curriculum vitae.