Computer Analysis of Architecture Using Automatic Image Understanding

Fan Wei,Yuan Li,Lior Shamir

doi:10.46298/jdmdh.4683

Abstract

In the past few years, computer vision and pattern recognition systems have been becoming increasingly more powerful, expanding the range of automatic tasks enabled by machine vision. Here we show that computer analysis of building images can perform quantitative analysis of architecture, and quantify similarities between city architectural styles in a quantitative fashion. Images of buildings from 18 cities and three countries were acquired using Google StreetView, and were used to train a machine vision system to automatically identify the location of the imaged building based on the image visual content. Experimental results show that the automatic computer analysis can automatically identify the geographical location of the StreetView image. More importantly, the algorithm was able to group the cities and countries and provide a phylogeny of the similarities between architectural styles as captured by StreetView images. These results demonstrate that computer vision and pattern recognition algorithms can perform the complex cognitive task of analyzing images of buildings, and can be used to measure and quantify visual similarities and differences between different styles of architectures. This experiment provides a new paradigm for studying architecture, based on a quantitative approach that can enhance the traditional manual observation and analysis. The source code used for the analysis is open and publicly available.

Highlights

Architecture is one of the oldest forms of the combination of science and art
While beauty and aesthetics are subjective concepts that vary between different cultures (Carlson, 2002), architecture in different eras and geographic locations is fundamentally different, and the differences are the result of the complex combination of social, cultural, climatic, historical, religious, and geological influences (Fletcher, 1931)
To first identify whether the algorithm is sensitive to architecture data, the accuracy of which a building image can be associated with a city was tested

Summary

INTRODUCTION

Architecture is one of the oldest forms of the combination of science and art. In addition to usability and environmental aspects of buildings, architecture has substantial aesthetic considerations, and the beauty of the building is considered one of the three most important aspects by which architecture is measured, along with the usability of the building and its durability (Vitruvius, 30BC). Other work focused on systems that can use such techniques for the purpose of preservation and cultural heritage (Merchán et al, 2018) Another task of using computer analysis of visual architectural data is the ranking and estimation of urban environment quality (Liu et al, 2018), profiling how urban environments are perceived (Dubey et al, 2016), measuring the livelihood of neighborhoods (De Natai et al, 2016), or automatic estimation of a building age (Zeppelzauer et al, 2018; Lee et al, 2015). The data used in this study are images of chosen cities and countries, all were collected using Google StreetView, which has been used in the past for automatic analysis of the changes in architectural style over time (Lee et al, 2015).

IMAGE ANALYSIS

SEPARATING REGIONS OF INTEREST

RESULTS

CONCLUSIONS