Author: Seyedeh Bahar Ahmadi

Publisher:

Published: 2021

Total Pages: 0

ISBN-13:

DOWNLOAD EBOOK

Microscopy is becoming one of the major practices for researching in any discipline, but new challenges are introduced by the immense quantities of raw data produced by its procedures. To obtain meaningful information, raw data must be visualized and processed. Storing and visualizing raw microscopy data, however, is very expensive. Visualization of 3D images is made possible by commercial software packages installed on high-powered computers. In addition, most of the time the data needs to be segmented and labeled, a process that is usually achieved through manual processes that are inherently labor-intensive and subject to error. Computer vision and machine learning algorithms hold great promise to change the paradigms of microscopy, but there are certain challenges that have prevented widespread use of data analytics in state-of-the-art microscopy. Currently practiced data analytics techniques are application-oriented and thus often cannot be utilized as universal solutions. Also, despite the great advances in automated data processing, an insightful interpretation of microscopy data still relies on manual investigation by subject matter experts. The complex and multidisciplinary nature of problems that involve microscopy data calls for a collaboration between multiple subject matter experts from different areas of expertise to provide complementary insights. An effective collaborative analysis of microscopy data therefore requires that researchers can investigate the data simultaneously. Furthermore, the fuel of any machine learning technique is training data, a great portion of which can be generated cognitively. This includes the opinions of the subject matter experts and the data produced through tasks that are trivial for humans but complex for machines (e.g. segmentation). In this work, we are presenting a novel platform for visualizing and collaborating on microscopy images without subjecting any cost on the user. Through this platform, researchers can investigate the data collaboratively and make assessments. Collecting these assessments and overlaying them on the other forms of data forms the basis of training data. This training data will in turn be used to empower machine learning techniques that will allow researchers to achieve greater insights more easily.

Download or read book Cloud-based Multi Modal Crowdsourcing Image Visualization and Machine Learning Enabled Image Analysis Platform written by Seyedeh Bahar Ahmadi and published by . This book was released on 2021 with total page 0 pages. Available in PDF, EPUB and Kindle. Book excerpt: Microscopy is becoming one of the major practices for researching in any discipline, but new challenges are introduced by the immense quantities of raw data produced by its procedures. To obtain meaningful information, raw data must be visualized and processed. Storing and visualizing raw microscopy data, however, is very expensive. Visualization of 3D images is made possible by commercial software packages installed on high-powered computers. In addition, most of the time the data needs to be segmented and labeled, a process that is usually achieved through manual processes that are inherently labor-intensive and subject to error. Computer vision and machine learning algorithms hold great promise to change the paradigms of microscopy, but there are certain challenges that have prevented widespread use of data analytics in state-of-the-art microscopy. Currently practiced data analytics techniques are application-oriented and thus often cannot be utilized as universal solutions. Also, despite the great advances in automated data processing, an insightful interpretation of microscopy data still relies on manual investigation by subject matter experts. The complex and multidisciplinary nature of problems that involve microscopy data calls for a collaboration between multiple subject matter experts from different areas of expertise to provide complementary insights. An effective collaborative analysis of microscopy data therefore requires that researchers can investigate the data simultaneously. Furthermore, the fuel of any machine learning technique is training data, a great portion of which can be generated cognitively. This includes the opinions of the subject matter experts and the data produced through tasks that are trivial for humans but complex for machines (e.g. segmentation). In this work, we are presenting a novel platform for visualizing and collaborating on microscopy images without subjecting any cost on the user. Through this platform, researchers can investigate the data collaboratively and make assessments. Collecting these assessments and overlaying them on the other forms of data forms the basis of training data. This training data will in turn be used to empower machine learning techniques that will allow researchers to achieve greater insights more easily.