Loading...

An Assessment of the Applications of the Classical and Modern Image Processing Techniques in Improving House Price Estimators

Salamirad, Amir Hossein | 2020

433 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 52759 (01)
  4. University: Sharif University of Technology
  5. Department: Industrial Engineering
  6. Advisor(s): Shadrokh, Shahram; Shavandi, Hassan
  7. Abstract:
  8. Since the advent of online real estate database companies like Zillow, Trulia and Redfin, the problem of automatic estimation of market values for houses has received considerable attention. Several real estate websites provide such estimates using a proprietary formula. Although these estimates are often close to the actual sale prices, in some cases they are highly inaccurate. One of the key factors that affects the value of a house is its interior and exterior appearance, which is not considered in calculating automatic value estimates. In this work, we evaluated the impact of visual characteristics of a house the buyer's proposed price. Using deep convolutional neural networks on a large dataset of photos of home interiors and exteriors (which we obtained from Zillow's website), we developed a method for estimating the best "Correction Factor" --which has to be multiplied in Zestimate Price so as to decrease the prediction error—from the house's images. We also develope a novel framework for automated value assessment using the above photos in addition to home characteristics including size, zip code, number of bedrooms, to name but a few. Finally, by applying our proposed method for price estimation to a Test set containing the images and metadata of 5000 houses, we demonstrated that this method reduces the MAE of the Zestimate's estimations
  9. Keywords:
  10. Computer Vision ; Deep Learning ; Machine Vision ; Housing Price ; Correction Factor ; Convolutional Neural Network ; Housing Pricing

 Digital Object List

 Bookmark

No TOC