ImmoAge - Visual Age Prediction of Real Estate

Year of construction, architectural period and architectural style have a significant impact on property prices. New methods of classifying and evaluating real estate have been developed on the basis of automated image analysis.

Age and value of real estate

While real estate appraisal based on hedonic regression analysis encompasses the assessment of different locations, there is currently no automatic method of classifying buildings according to their age. The ImmoAge project is developing new image analysis methods for the purpose of extracting age data from images. The project aims to develop a method of data mining buildings’ age- and location-specific visual patterns as well as automatically classifying buildings by age and location.

Automated classification of real estate

Developing a new method of image analysis requires large amounts of data in order to ensure accurate identification of visual patterns for different architectural periods or styles. The project consortium is providing the necessary amount of real estate assessments and reviews including visual material and metadata, which can be used for deeper automated analysis.
Using this material, we will compile a comprehensive data set for further exploration of visual data mining methods and machine learning.

New methods of analysis

The main focus of the ImmoAge project is developing existing methods of image analysis and automated classification of real estate properties in order to extract specific information about location, year of construction, architectural period and architectural style. The project represents a first step in the joint ImmoPixel project to generate novel real estate benchmarking tools.

Publications

Download Bibtex for all results

Koch, D., Despotovic, M., Thaler, S., & Zeppelzauer, M. (2021). Where do University Graduates live? – A Computer Vision Approach using Satellite Images. Applied Intelligence, 51, 8088–8105. https://doi.org/https://doi.org/10.1007/s10489-021-02268-8

Koch, D., Despotovic, M., Döller, M., Leiber, S., & Zeppelzauer, M. (2020). Computer Vision in Building Research: An Application for Prediction of Condition and Costs of a Property. Building Research & Information, Submitted.

Bibtex

Bibtex

Koch, D., Despotovic, M., Sascha, L., Sakeena, M., Döller, M., & Zeppelzauer, M. (2020). Real Estate Image Analysis - A Literature Review. Journal of Real Estate Literature, 27(2), 269–300. https://doi.org/10/gnt2wg

Despotovic, M., Koch, D., Leiber, S., Döller, M., Sakeena, M., & Zeppelzauer, M. (2019). Prediction and analysis of heating energy demand for detached houses by computer vision. Energy & Buildings, 193, 29–35. https://doi.org/10/fsxn

Zeppelzauer, M. (2018). Visual Estimation of Building Condition with Patch-level ConvNets. http://dl.acm.org/citation.cfm?doid=3210499.3210526

Koch, D., Despotovic, M., Sakeena, M., Döller, M., & Zeppelzauer, M. (2018). Visual Estimation of Building Condition with Patch-level ConvNets. Proceedings of the 2018 ACM Workshop on Multimedia for Real Estate Tech - RETech"18, 12–17. https://doi.org/10/ghpp2m

Zeppelzauer, M., Despotovic, M., Sakeena, M., Koch, D., & Döller, M. (2018). Automatic Prediction of Building Age from Photographs. Proceedings of the ACM International Conference on Multimedia Retrieval (ICMR "18), 126–134. https://doi.org/10/ghpp2k

Despotovic, M., Sakeena, M., Koch, D., Döller, M., & Zeppelzauer, M. (2017). Predicting Heating Energy Demand by Computer Vision. Computer Science - Research and Development, 33, 231–232. https://doi.org/10/gh3772