BEGIN:VCALENDAR VERSION:2.0 PRODID:-//https://techplay.jp//JP CALSCALE:GREGORIAN METHOD:PUBLISH X-WR-CALDESC:[22nd AIP Open Seminar] Talks by Machine Intelligence for Medi cal Engineering Team X-WR-CALNAME:[22nd AIP Open Seminar] Talks by Machine Intelligence for Medi cal Engineering Team X-WR-TIMEZONE:Asia/Tokyo BEGIN:VTIMEZONE TZID:Asia/Tokyo BEGIN:STANDARD DTSTART:19700101T000000 TZOFFSETFROM:+0900 TZOFFSETTO:+0900 TZNAME:JST END:STANDARD END:VTIMEZONE BEGIN:VEVENT UID:813023@techplay.jp SUMMARY:[22nd AIP Open Seminar] Talks by Machine Intelligence for Medical E ngineering Team DTSTART;TZID=Asia/Tokyo:20210421T150000 DTEND;TZID=Asia/Tokyo:20210421T170000 DTSTAMP:20260508T055159Z CREATED:20210329T060014Z DESCRIPTION:イベント詳細はこちら\nhttps://techplay.jp/event/81302 3?utm_medium=referral&utm_source=ics&utm_campaign=ics\n\nMachine Intellig ence for Medical Engineering Team (https://aip.riken.jp/labs/goalorient_t ech/machine_intell_med_eng/) at RIKEN AIP\n\nSpeaker 1: Tatsuya Harada\nT itle: Overview of Machine Intelligence for Medical Engineering Team\nAbst ract: \nWe will present an overview of the Machine Intelligence for Medic al Engineering Team. Medical information processing requires handling mul timodal information such as 3D volumes\, medical records with various con straints\, and privacy-aware images. To tackle these topics\, we have dev eloped fundamental ML-based methods for 3D information processing\, deep neural networks for tree structures\, and privacy-aware knowledge transfe r. We will also introduce those methods in this talk. \n\nSpeaker 2: Yusu ke Mukuta\nTitle: Invariant Feature Coding using Tensor Product Represent ation\nAbstract: \nExploiting the invariance is important for the efficie nt feature learning. We propose a method that incorporates the invariance into the image feature coding\, which is the method to use the statistic s of the local features as the global image feature. To this end\, we reg ard the existing feature coding function as the tensor product of the loc al feature functions and then calculate its invariance as the global feat ure. The proposed method demonstrates better classification accuracy with robustness to the considered transformation using the smaller global fea ture dimension on several image recognition datasets including medical im age analysis.\n\nSpeaker 3: Lin Gu\nTitle: Limited Data and Interpretabil ity：the challenge and solution for Real-world Medical AI\nAbstract: \nT hough deep learning has shown successful performance in the medical image analysis in the tasks of classifying the label and severity stage of cer tain disease. However\, the CNN based methods suffer the bottleneck of l acking training label and interpretability. For example\, most of them gi ve few evidence on how to make prediction. To make it worse\, the ubiquit ous adversarial attack has posed even more serious challenge on its real application. This talk would introduce our recent progress for these chal lenges on various medical image domains for real-world application.\n\nSp eaker 4: Yusuke Kurose\nTitle: Machine Learning for Medical Image Diagnos is\nAbstract: \nThe development of machine learning in recent years has b een remarkable and has influenced many fields\, and it is no exception in medical image processing. However\, there are unique problems with medic al image processing. For example\, a very high resolution on a pathologic al image. It is not able to be applied to a general segmentation method d irectly due to the very high resolution on it. To solve this problem\, a general method for pathological tissue classification uses a small patch extracted from the image as a classification input. However\, it cannot c onsider a global feature in the tissue for classification. In this presen tation\, I will introduce our pathological segmentation method which can consider the global feature for the very high-resolution image. Also\, I introduce other problems on medical image processing and our developed me thods to solve them. LOCATION:オンライン URL:https://techplay.jp/event/813023?utm_medium=referral&utm_source=ics&utm _campaign=ics END:VEVENT END:VCALENDAR