City Forensics: Using Visual Elements to Predict Non-Visual City Attributes

Arietta, Sean M.; Efros, Alexei A.; Ramamoorthi, Ravi; Agrawala, Maneesh

doi:10.1109/tvcg.2014.2346446

Cited by 122 publications

(66 citation statements)

References 41 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Been et al (20) Moreover, crowdsourcing and computer vision methods have been used along with street-level imagery to identify geographically distinctive architectural elements (23), develop unique city signatures (24), and predict socioeconomic indicators (25,26). Taken together, the range of findings illustrates how computer vision methods can be used to improve the quantitative study of urban appearance and space.…”

Section: Significancementioning

confidence: 99%

Computer vision uncovers predictors of physical urban change

Naik

Kominers

Raskar

et al. 2017

Proc. Natl. Acad. Sci. U.S.A.

218

132

View full text Add to dashboard Cite

Which neighborhoods experience physical improvements? In this paper, we introduce a computer vision method to measure changes in the physical appearances of neighborhoods from time-series street-level imagery. We connect changes in the physical appearance of five US cities with economic and demographic data and find three factors that predict neighborhood improvement. First, neighborhoods that are densely populated by college-educated adults are more likely to experience physical improvements-an observation that is compatible with the economic literature linking human capital and local success. Second, neighborhoods with better initial appearances experience, on average, larger positive improvements-an observation that is consistent with "tipping" theories of urban change. Third, neighborhood improvement correlates positively with physical proximity to the central business district and to other physically attractive neighborhoods-an observation that is consistent with the "invasion" theories of urban sociology. Together, our results provide support for three classical theories of urban change and illustrate the value of using computer vision methods and street-level imagery to understand the physical dynamics of cities.urban economics | gentrification | urban studies | computer vision | neighborhood effects F or more than a century, urban planners, economists, sociologists, and architects have advanced theories connecting the dynamics of a neighborhood's physical appearance to its location, demographics, and built infrastructure.The tipping theory of Schelling (1) and Grodzins (2) suggests that neighborhoods in bad physical condition will get progressively worse, whereas nicer areas will get better. Economic theories of urban change at the city level often emphasize population density and education (3-6), and it is natural to hypothesize that agglomeration of human capital will predict neighborhood-level improvements as well. Theories from urban sociology, such as the invasion theory of Burgess (7), however, emphasize locations and social networks, predicting that improvements in a city's appearance should be spatially clustered, and that improvements should occur both near the central business districts (CBDs) and near other physically attractive neighborhoods.To test theories of physical neighborhood change, we need to quantify neighborhood appearance at different points in time. Historically, however, methods to quantify neighborhood appearance have not been scalable. The empirical literature on urban appearance, which was pioneered by urban planners such as Lynch (8), Rapoport (9), and Nasar (10), as well as by psychologists such as Milgram (11), has relied on interviews, lowthroughput visual perception surveys, and manual evaluation of images. Those methods, however, can only be used to collect data on a few neighborhoods and have limited spatial resolution. In the past decade, new data on urban appearance have emerged in the form of "street view" imagery (12). As of 2016, Google Street View has photographe...

show abstract

Section: Significancementioning

confidence: 99%

Computer vision uncovers predictors of physical urban change

Naik

Kominers

Raskar

et al. 2017

Proc. Natl. Acad. Sci. U.S.A.

218

132

View full text Add to dashboard Cite

show abstract

“…Urban Perception and High-Definition Mapping Recently there has been a surge in interest for applying techniques for scene classification [2,4,17,19,22,27,28] and image segmentation to understanding urban areas and transportation infrastructure [5,16,25]. The former focuses on higher-level labels, such as perceived safety, population density, or beauty, while the later focuses on finer-grained labels, such as the location on line markings or the presence of sidewalks.…”

Section: Scene Classification and Image Segmentationmentioning

confidence: 99%

FARSA: Fully Automated Roadway Safety Assessment

Song

Workman

Hadžić

et al. 2018

2018 IEEE Winter Conference on Applications of Computer Vision (WACV)

View full text Add to dashboard Cite

This paper addresses the task of road safety assessment. An emerging approach for conducting such assessments in the United States is through the US Road Assessment Program (usRAP), which rates roads from highest risk (1 star) to lowest (5 stars). Obtaining these ratings requires manual, fine-grained labeling of roadway features in streetlevel panoramas, a slow and costly process. We propose to automate this process using a deep convolutional neural network that directly estimates the star rating from a street-level panorama, requiring milliseconds per image at test time. Our network also estimates many other roadlevel attributes, including curvature, roadside hazards, and the type of median. To support this, we incorporate taskspecific attention layers so the network can focus on the panorama regions that are most useful for a particular task. We evaluated our approach on a large dataset of real-world images from two US states. We found that incorporating additional tasks, and using a semi-supervised training approach, significantly reduced overfitting problems, allowed us to optimize more layers of the network, and resulted in higher accuracy.

show abstract

“…The challenge is that ground-level imagery is sparsely distributed, especially away from major urban areas and tourist attractions. This means that methods which estimate maps using only ground-level imagery [1,21,36] typically generate either low-resolution or noisy maps.…”

Section: Mapping Image Scenicnessmentioning

confidence: 99%

Understanding and Mapping Natural Beauty

Workman

Souvenir

Jacobs

2017

2017 IEEE International Conference on Computer Vision (ICCV)

View full text Add to dashboard Cite

While natural beauty is often considered a subjective property of images, in this paper, we take an objective approach and provide methods for quantifying and predicting the scenicness of an image. Using a dataset containing hundreds of thousands of outdoor images captured throughout Great Britain with crowdsourced ratings of natural beauty, we propose an approach to predict scenicness which explicitly accounts for the variance of human ratings. We demonstrate that quantitative measures of scenicness can benefit semantic image understanding, content-aware image processing, and a novel application of cross-view mapping, where the sparsity of ground-level images can be addressed by incorporating unlabeled overhead images in the training and prediction steps. For each application, our methods for scenicness prediction result in quantitative and qualitative improvements over baseline approaches.

show abstract

City Forensics: Using Visual Elements to Predict Non-Visual City Attributes

Cited by 122 publications

References 41 publications

Computer vision uncovers predictors of physical urban change

Computer vision uncovers predictors of physical urban change

FARSA: Fully Automated Roadway Safety Assessment

Understanding and Mapping Natural Beauty

Contact Info

Product

Resources

About