Conference Presentations by Ishan Saksena
The majority of current automated algorithms for estimating house prices only take into account t... more The majority of current automated algorithms for estimating house prices only take into account textual information like the home's neighbourhood and the number of rooms. A human representative makes a physical inspection of the home and makes an estimate of the actual cost. In this work, we suggest merging written information about the property with visual elements extracted from images of the house. The combined features are fed into a Convolution Neural Network (NN), which produces a single output that is an estimate of the price of the home. For the purpose of developing and evaluating our network, we have gathered a housing dataset that includes both visual and textual features. The dataset consists of the price and other features of 500 homes from Bangalore along with images of those homes.
Uploads
Conference Presentations by Ishan Saksena