Close collaboration is possible even over a long distance\u00a0<\/em><\/h3>\n<\/blockquote>\n <\/p>\n\n\t\t
\n\t\t\t\n\t\t\t\t\n\t\t\t
\n\t\t\t<\/div>\n\t\t<\/div> \n\t\t\n\t\t\tToday Arne Schmidt (ESR8) talks about his virtual secondment at the University of Stavanger (UiS) about Whole Slide Image (WSI) preprocessing, Multiple Instance Learning with attention and the classification of bladder cancer.<\/span>\n\t\t<\/div>\n\t\t\t<\/div>\n\t\t<\/div>\n <\/p>\n
Due to the pandemic, traveling to Norway was not possible, but that did not stop the\u00a0UiS\u00a0from hosting the secondment of ESR8 \u2013 virtually! Supervised by Trygve\u00a0Eftest\u00f8l\u00a0and Kjersti Engan, Arne Schmidt (ESR8) was the virtual\u00a0vistor\u00a0at the\u00a0UiS\u00a0from April until July 2021.<\/span>\u00a0<\/span><\/p>\nThe main topic of this secondment was the preprocessing of WSIs for the automatic detection of cancer.\u00a0<\/span>The preprocessing of the data is essential to successfully train a machine learning algorithm: it prevents the final dataset from containing unwanted artifacts (such as burnt or folded tissue) and unnecessary information (background,\u00a0unrelevant\u00a0tissue regions).\u00a0Additionally\u00a0it ensures the generalizability of the data through color normalization or image augmentation. Apart from informative meetings and seminars, a collaboration was established to gather common practices and the state of the art in literature. This joint work resulted in the Deliverable D3.1 “Preprocessing and Standardization Protocol” of the CLARIFY project. The main findings will be published in a research article that is currently written by Neel Kanwal (UiS, ESR4), Fernando P\u00e9rez Bueno (UGR) and Arne Schmidt (ESR8, UGR).<\/span>\u00a0<\/span><\/p>\nAnother topic of fruitful collaboration was the Multiple Instance Learning (MIL) with attention mechanisms. Here the goal is to train a machine learning model with only WSI labels instead of detailed annotations. As this topic is a major research interest to both Saul Fuster Navarro (ESR5,\u00a0UiS) and Arne Schmidt (ES8, UGR), several meetings were held to exchange experiences and ideas.<\/span>\u00a0<\/span><\/p>\nLast but not least, ESR8 could gain some insights about the classification of bladder cancer with artificial intelligence. At the\u00a0UiS\u00a0some researchers (such as Rune Wetteland) have several years of experience in this topic and gave some informative seminars about this topic. In the future, joint work on the classification of bladder cancer with MIL attention\u00a0is planned to further extend this international cooperation within the CLARIFY project.<\/span>\u00a0<\/span><\/p>\n
\n\t\t\t<\/p>\n
Due to the pandemic, traveling to Norway was not possible, but that did not stop the\u00a0UiS\u00a0from hosting the secondment of ESR8 \u2013 virtually! Supervised by Trygve\u00a0Eftest\u00f8l\u00a0and Kjersti Engan, Arne Schmidt (ESR8) was the virtual\u00a0vistor\u00a0at the\u00a0UiS\u00a0from April until July 2021.<\/span>\u00a0<\/span><\/p>\n The main topic of this secondment was the preprocessing of WSIs for the automatic detection of cancer.\u00a0<\/span>The preprocessing of the data is essential to successfully train a machine learning algorithm: it prevents the final dataset from containing unwanted artifacts (such as burnt or folded tissue) and unnecessary information (background,\u00a0unrelevant\u00a0tissue regions).\u00a0Additionally\u00a0it ensures the generalizability of the data through color normalization or image augmentation. Apart from informative meetings and seminars, a collaboration was established to gather common practices and the state of the art in literature. This joint work resulted in the Deliverable D3.1 “Preprocessing and Standardization Protocol” of the CLARIFY project. The main findings will be published in a research article that is currently written by Neel Kanwal (UiS, ESR4), Fernando P\u00e9rez Bueno (UGR) and Arne Schmidt (ESR8, UGR).<\/span>\u00a0<\/span><\/p>\n Another topic of fruitful collaboration was the Multiple Instance Learning (MIL) with attention mechanisms. Here the goal is to train a machine learning model with only WSI labels instead of detailed annotations. As this topic is a major research interest to both Saul Fuster Navarro (ESR5,\u00a0UiS) and Arne Schmidt (ES8, UGR), several meetings were held to exchange experiences and ideas.<\/span>\u00a0<\/span><\/p>\n Last but not least, ESR8 could gain some insights about the classification of bladder cancer with artificial intelligence. At the\u00a0UiS\u00a0some researchers (such as Rune Wetteland) have several years of experience in this topic and gave some informative seminars about this topic. In the future, joint work on the classification of bladder cancer with MIL attention\u00a0is planned to further extend this international cooperation within the CLARIFY project.<\/span>\u00a0<\/span><\/p>\n
![\"\"](\"https:\/\/cvblab.webs.upv.es\/clarify-project.eu\/wp-content\/uploads\/2021\/09\/Preprocessing_pipeline.png\")
<\/span><\/span><\/span><\/a><\/p>\n