Sano PhD student Monika Pytlarz presented her work at “SPIE Medical Imaging” conference in San Diego, USA
Monika presented a scientific poster entitled “Automated glioma multiclass tumor classification” during a poster session of the “SPIE Medical Imaging” conference on 18 –24th February 2023. The poster demonstrated an automatic deep-learning method for grading glioma malignant tumors given a small imbalanced dataset of cores of stained tissue microarrays. This work is a result of cooperation with the Nencki Institute of Experimental Biology PAS. The poster inspired some interesting discussions about the advantages of this histological imaging technique, different data augmentation methods, and architectures suitable for image classification.
The conference gave Monika a great opportunity to meet outstanding scientists. According to Monika’s report, one of the most impressive presentations was made by Professor Elizabeth Krupinski on the topic of translating computational innovations into reality and focusing on the users during this process. Prof. Krupinski proved by data the importance to give radiologists (and pathologists) tools for decision support, for quantification of image features, for automatic generation of radiology reports, and for auto reproduction of those to patient-friendly deconvoluted form.
Monika spent most of the time on digital pathology presentation sessions where inspiring approaches to solving the computational pathology challenges were shared, and also participated in an exciting course offered by the Fraunhofer-Institut on vision transformers, which recently entered the scene of medical image analysis.