ImmoAge

#Research #Institute of Creative\Media/Technologies #Media Computing #National third-party funding #National co-operation project

Year and period of construction or style have a significant influence on real estate value. Automated image analysis provides new methods to classify and evaluate real estate property.

  • Project manager:
    FH-Prof. Dipl.-Ing. Mag. Dr. Matthias Zeppelzauer
  • External Staff:
    David Koch, University of Applied Sciences Kufstein
    Miroslav Despotovic, University of Applied Sciences Kufstein
  • Partner:
    University of Applied Sciences Kufstein
    R&S Software GmbH
  • Weblinks:
    http://mc.fhstp.ac.at/projects/2017/immoage
  • Funding:
    Austrian Research Promotion Agency
  • Runtime:
    01.10.201630.09.2018
  • Category:
    National third-party funding
    National co-operation project
  • Projectstatus:
    ongoing

Age and Value of Real Estate


While real estate appraisal, an analysis based on hedonic regression, includes the assessment of different locations, automatic classification of buildings by their age remains unsolved. ImmoAge develops new image analysis methods to extract said information from a data set of images. The project aims at establishing data mining of age- and location-specific visual patterns of buildings as well as automated classification of those buildings by age and location.

Automated classification of real estate


To develop a new method of image analysis large amounts of data are necessary in order to ensure a solid identification of visual patterns for different periods or style of construction. The project consortium provides a corresponding amount of real estate assessments and reviews including visual material and metadata, which offer the opportunity for deeper automated analysis.
Based on this material we will compile a comprehensive data set to further explore visual data mining methods and machine learning.

New methods of analysis


“ImmoAge” centers on the further development of image analysis and automated classification of real estate properties to extract specific information about location, year and period of construction and style. The project represents a first step in our joint effort “ImmoPixel” to generate novel real estate benchmarking tools.