Enabling Video Privacy through Computer Vision

Senior, Andrew W.; Pankanti, Sharath; Hampapur, Arun; Brown, Lisa M.; Tian, Yingli; Ekin, Ahmet; Connell, Jonathan H.; Shu, Chiao-Fe; Lu, Max

doi:10.1109/msp.2005.65

Cited by 210 publications

(158 citation statements)

References 4 publications

Supporting

Mentioning

151

Contrasting

Unclassified

Order By: Relevance

“…Chen et al built a system for people detection, tracking and recognition [11] using gait analysis, face detection and face recognition. As visual privacy is a significant concern of such systems, a number of researchers have explored how to integrate privacy systems into these surveillance systems [48,56,12,10,18]. Others such as Chinomi et al have evaluated different methods for obscuring people in video data [13].…”

Section: Related Workmentioning

confidence: 99%

Respectful Cameras: Detecting Visual Markers in Real-Time to Address Privacy Concerns

Schiff

Meingast

Mulligan

et al. 2009

Protecting Privacy in Video Surveillance

View full text Add to dashboard Cite

To address privacy concerns regarding digital video surveillance cameras, we propose a practical, real-time approach that preserves the ability to observe actions while obscuring individual identities. In the Respectful Cameras system, people who wish to remain anonymous wear colored markers such as hats or vests. The system automatically tracks these markers using statistical learning and classification to infer the location and size of each face. It obscures faces with solid ellipsoidal overlays, while minimizing the overlay area to maximize the remaining observable region of the scene. Our approach uses a visual color-tracker based on a nine dimensional color-space using a Probabilistic Adaptive Boosting (AdaBoost) classifier with axis-aligned hyperplanes as weak hypotheses. We then use Sampling Importance Resampling (SIR) Particle Filtering to incorporate interframe temporal information. Because our system explicitly tracks markers, our system is well-suited for applications with dynamic backgrounds or where the camera can move (e.g. under remote control). We present experiments illustrating the performance of our system in both indoor and outdoor settings, with occlusions, multiple crossing targets, lighting changes, and observation by a moving robotic camera. Results suggest that our implementation can track markers and keep false negative rates below 2%. Fig. 1 A sample video frame is on left. The system has been trained to track green vests such as the one worn by the man with the outstretched arm. The system output is shown in the frame on the right, where an elliptical overlay hides the face of this man. The remainder of the scene including faces of workers not wearing green vests, remain visible. Note how the system successfully covers the face even when the vest is subjected to a shadow and a partial occlusion. Please visit "http://goldberg.berkeley.edu/RespectfulCameras" for more examples including video sequences.

show abstract

Section: Related Workmentioning

confidence: 99%

Respectful Cameras: Detecting Visual Markers in Real-Time to Address Privacy Concerns

Schiff

Meingast

Mulligan

et al. 2009

Protecting Privacy in Video Surveillance

View full text Add to dashboard Cite

show abstract

“…The system introduced in [3] relies on computer vision to analyze the video content. Depending on users' accesscontrol rights, different versions of the video are then presented where privacy-sensitive information is removed.…”

Section: Introductionmentioning

confidence: 99%

A framework for the validation of privacy protection solutions in video surveillance

Dufaux

Ebrahimi

2010

2010 IEEE International Conference on Multimedia and Expo

View full text Add to dashboard Cite

The issue of privacy protection in video surveillance has drawn a lot of interest lately. However, thorough performance analysis and validation is still lacking, especially regarding the fulfillment of privacy-related requirements. In this paper, we put forward a framework to assess the capacity of privacy protection solutions to hide distinguishing facial information and to conceal identity. We then conduct rigorous experiments to evaluate the performance of face recognition algorithms applied to images altered by privacy protection techniques. Results show the ineffectiveness of naïve privacy protection techniques such as pixelization and blur. Conversely, they demonstrate the effectiveness of more sophisticated scrambling techniques to foil face recognition.

show abstract

“…Senior et al [12] note that if the human or computer on the receiving side of the obfuscated video has access to references images (e.g. non-obfuscated images of the video subjects), then care must be taken to ensure that transmitted images do not reveal information which can be used to match to a reference image (using techniques such as face, gait or scene recognition).…”

mentioning

confidence: 99%

Integrating Utility into Face De-identification

Gross

Airoldi

Malin

et al. 2006

Lecture Notes in Computer Science

124

116

View full text Add to dashboard Cite

Abstract. With the proliferation of inexpensive video surveillance and face recognition technologies, it is increasingly possible to track and match people as they move through public spaces. To protect the privacy of subjects visible in video sequences, prior research suggests using ad hoc obfuscation methods, such as blurring or pixelation of the face. However, there has been little investigation into how obfuscation influences the usability of images, such as for classification tasks. In this paper, we demonstrate that at high obfuscation levels, ad hoc methods fail to preserve utility for various tasks, whereas at low obfuscation levels, they fail to prevent recognition. To overcome the implied tradeoff between privacy and utility, we introduce a new algorithm, k -Same-Select, which is a formal privacy protection schema based on k -anonymity that provably protects privacy and preserves data utility. We empirically validate our findings through evaluations on the FERET database, a large real world dataset of facial images.

show abstract

Enabling Video Privacy through Computer Vision

Cited by 210 publications

References 4 publications

Respectful Cameras: Detecting Visual Markers in Real-Time to Address Privacy Concerns

Respectful Cameras: Detecting Visual Markers in Real-Time to Address Privacy Concerns

A framework for the validation of privacy protection solutions in video surveillance

Integrating Utility into Face De-identification

Contact Info

Product

Resources

About