• Homepage
  • mmc2021 Abstract Database
  • Visual Data Analysis: Bringing the Human into the Loop
  • Visual Data Analysis: Bringing the Human into the Loop

    Abstract number
    140
    DOI
    10.22443/rms.mmc2021.140
    Corresponding Email
    [email protected]
    Session
    Stream 3: 3D+ Image Analysis
    Authors
    Dr. rer. nat. Daniel Baum (1)
    Affiliations
    1. Zuse Institute Berlin
    Keywords

    visual data analysis, image analysis, feature extraction, feature tracking

    Abstract text

    The amount of data produced every day is ever increasing. This is also true for all kinds of image data that are nowadays being acquired in almost every scientific field. Due to the overwhelming amount of data, there has been a focus on automating data analysis in recent years, spurred by the tremendous advances in artificial intelligence. However, some data analysis tasks remain too challenging to be performed fully automatically. These tasks can often only get solved by allowing some user input at various stages of the analysis workflow. Adding user interaction to support the analysis task also often allows the generalization of the analysis tools to entirely different applications. This generalization can be further boosted by modularizing the workflow such that parts of it can be easily replaced by data- and application-specific automatic or interactive solutions.

    This talk will show some image analysis examples that require the human-in-the-loop to analyze the data, which might be acquired using very different imaging techniques such as electron tomography, 2-photon microscopy, and micro-computed tomography. We will present solutions that work for both biological data as well as data from materials science. Even though the development of our solutions is always driven by particular applications, abstracting from the actual application data often leads to more general and more flexible solutions.