Large-scale privacy protection in Google Street View

Frome, Andrea; Cheung, German; Abdulkader, Ahmad; Zennaro, Marco; Wu, Bo; Bissacco, Alessandro; Adam, Hartwig; Neven, Hartmut; Vincent, Luc

doi:10.1109/iccv.2009.5459413

Cited by 206 publications

(131 citation statements)

References 12 publications

Supporting

Mentioning

130

Contrasting

Unclassified

Order By: Relevance

“…The growing ubiquity of mobile imaging devices has led to applications in various environments, for example, autonomous navigation in urban settings, vision based indoor assistance for mobile robots [13], and so on. In particular, the proliferation of online databases has led to the necessity to recognize and obscure text in street level imagery for privacy protection [14].…”

Section: Introductionmentioning

confidence: 99%

Exemplar Driven Character Recognition in the Wild

Sheshadri¹,

Divvala²

2012

Procedings of the British Machine Vision Conference 2012

View full text Add to dashboard Cite

Character recognition in natural scenes continues to represent a formidable challenge in computer vision. Beyond variation in font, there exist difficulties in occlusion, background clutter, binarisation, and arbitrary skew. Recent advances have leveraged state of the art methods from generic object recognition to address some of these challenges. In this paper, we extend the focus to Indic script languages (e.g., Kannada) that contain large character sets (order of 1000 classes unlike 62 in English) with very low inter character variation. We identify this scenario as a fine grained visual categorization task, and present a simple exemplar based multi layered classification approach to the problem. The proposed approach beats the existing state of the art on the chars74k and ICDAR datasets by over 10% for English, and around 24% for Kannada.

show abstract

Section: Introductionmentioning

confidence: 99%

Exemplar Driven Character Recognition in the Wild

Sheshadri¹,

Divvala²

2012

Procedings of the British Machine Vision Conference 2012

View full text Add to dashboard Cite

show abstract

“…From a practical perspective, storm drain inlet locations could be collected as part of drainage maintenance work or other regular work that is carried out on roads. Alternatively, an automated process could be developed allowing for a detection of storm drain inlet locations based on photo-interpretation of the application of Google Street View images, similar to algorithms that are applied for face or license plate recognition [40].…”

Section: Discussionmentioning

confidence: 99%

Urban Flood Simulation Using Synthetic Storm Drain Networks

Bertsch

Glenis

Kilsby

2017

Water

View full text Add to dashboard Cite

Recent developments in urban drainage modelling allow for a more realistic coupling of the two-dimensional (2D) surface and one-dimensional (1D) sub-surface drainage domain exchanging water through storm drain inlets instead of a sub-catchment approach based on manholes. Experience has shown, however, that comprehensive records of storm drain inlet locations are often missing or incomplete, preventing users accessing the full benefit of these modelling capabilities. Therefore, this study developed a GIS routine to generate synthetic storm drain inlet locations for the purpose of urban flood modelling. Hydrodynamic model results for a synthetically generated and surveyed storm drain inlet network were obtained using the CityCAT 1D/2D system. On a catchment scale the flow field (surface and flow captured by inlets) simulated by the network of synthetic storm drainage inlets shows satisfactory results when compared with that simulated using the actual network. The results also highlight the sensitivity of the inflows to relatively small changes in terms of the location of storm drain inlets and the effectiveness of storm drain inlets in ponding areas.

show abstract

“…One can easily spend days to weeks tweaking parameters. Another drawback of Viola and Jones' method is that only greyscale images are processed as opposed to recent approaches that have shown that color information may improve face detection results ( [45]). To overcome these drawbacks we present two major adaptions to the OpenCV system: Firstly, we introduce new feature layers that can be used either for color channels or arbitrary descriptors and secondly, we parallelize the complete training procedure to run on many-core architectures.…”

Section: Enhanced Adaboost Trainingmentioning

confidence: 99%

Fast Face Detector Training Using Tailored Views

Scherbaum

Petterson

Feris

et al. 2013

2013 IEEE International Conference on Computer Vision

View full text Add to dashboard Cite

Face detection is an important task in computer vision and often serves as the first step for a variety of applications. State-of-the-art approaches use efficient learning algorithms and train on large amounts of manually labeled imagery. Acquiring appropriate training images, however, is very time-consuming and does not guarantee that the collected training data is representative in terms of data variability. Moreover, available data sets are often acquired under controlled settings, restricting, for example, scene illumination or 3D head pose to a narrow range. This paper takes a look into the automated generation of adaptive training samples from a 3D morphable face model. Using statistical insights, the tailored training data guarantees full data variability and is enriched by arbitrary facial attributes such as age or body weight. Moreover, it can automatically adapt to environmental constraints, such as illumination or viewing angle of recorded video footage from surveillance cameras. We use the tailored imagery to train a new many-core implementation of Viola Jones' AdaBoost object detection framework. The new implementation is not only faster but also enables the use of multiple feature channels such as color features at training time. In our experiments we trained seven view-dependent face detectors and evaluate these on the Face Detection Data Set and Benchmark (FDDB). Our experiments show that the use of tailored training imagery outperforms state-of-the-art approaches on this challenging dataset.

show abstract

Large-scale privacy protection in Google Street View

Cited by 206 publications

References 12 publications

Exemplar Driven Character Recognition in the Wild

Exemplar Driven Character Recognition in the Wild

Urban Flood Simulation Using Synthetic Storm Drain Networks

Fast Face Detector Training Using Tailored Views

Contact Info

Product

Resources

About