Computer Vision News - March 2022

28 Medical Imaging Research I hope everyone had an amazing time last month and tried out some coding examples from our last issue! As always, feel free to reach out if you havemore ideas, examples, requests. Thank you kindly for sending always nice words and suggestions for our future issues. Let always be kind to people around us, educate and be patient! Keep up the amazing work you are doing in your life, professional and academic world, and most of all: enjoy it  Review This month the review article is a very relevant deep learning framework paper, called NaroNet: Discovery of tumor microenvironment elements from highly multiplexed images , published in Medical Image Analysis by Elsevier and authored by Daniel Jiménez- Sánchez, Mikel Ariz, Hang Chang, Xavier Matias-Guiu, Carlos E. de Andrea and Carlos Ortiz-de-Solórzano . We will try to do our best to review and explain to you in much less detail every article we find interesting from a scientific or exploratory perspective. If you want to learn, go deeper, and understand, even more, remember to always check the original references, article, and have a good discussion about it. Feel free to share our issue with your friends as well, as this may help evolve to new collaborations, conversations, and disputes (scientific disputes are always the best!) Introduction The histopathology and phenotype of a tumor guide its diagnosis, prognosis, and help to predict its response to anticancer treatments. Automating these tasks using machine learning (ML) is the goal of a novel field known as computational pathology. For instance, WSDL has been effectively used for tumor subtyping. SCA emerged in the context of the research for novel cancer biomarkers. SCA methods build topological networks containing cell phenotype interactions. They apply graph-based clustering to assign groups of cells to different neighborhoods. Since SCA methods use the cell as the basic unit of tissue representation, they provide a high level of interpretability. Naronet: Discovery of Tumor Microenvironment Elements from Highly Multiplexed Images