An accurate diagnosis of cancer often requires a histological examination, during which histopathologists analyze tissue samples at microscopic level. They aim not only at detecting cancer but also characterizing it. Cells’ size and shape, biomarkers, along with other elements are studied and contribute to defining the cancer type and grade. From this diagnosis, oncologists will decide on a treatment.
In histology, obtaining quality annotation is an expensive and lengthy process as it requires expert knowledge. Additionally, even for experts, giving a cell level precision on annotations is almost impossible and often subjective. Therefore, we speculate that being able to have slide level annotations or taking into account the annotation imprecision in our learning tasks could greatly reduce the need for annotation and improve models accuracy in lesion segmentation tasks. As a matter of fact, many authors already leverage the power of weakly supervised learning for histopathology [1, 2, 3].
During this internship, we will apply weakly supervised learning methods to the previously mentioned challenges. More precisely, the intern will perform:
[1] HistoSegNet: Semantic Segmentation of Histological Tissue Type in Whole Slide Images, Chan_2019
[2] A Comprehensive Analysis of Weakly-Supervised Semantic Segmentation in Different Image Domains, Chan, 2020
[3] Weakly supervised multiple instance learning histopathological tumor segmentation, Le Rousseau, 2021
Applied mathematics student.
Meeting with two data team members
These companies are also recruiting for the position of “Data / Business Intelligence”.