Searching surveillance video

Hampapur, Arun; Brown, Lisa M.; Feris, Rogério; Senior, A.; Shu, Chiao-Fe; Tian, Yingli; Zhai, Yun; Lu, Max

doi:10.1109/avss.2007.4425289

Cited by 21 publications

(9 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Thumbnails of the detected vehicles are displayed, and the user can click on them to view a video clip of the selected vehicle. The framework and software architecture of our vehicle search system is similar to the IBM Smart Surveillance Solution [8]. The search interface and a demonstration of our system in operation can be seen in the following video:…”

Section: Attribute Extraction / Searchmentioning

confidence: 99%

“…Traditional surveillance systems based on background modeling [17,8] generally fail to segment vehicles in crowds, as multiple vehicles tend to get clustered into a single motion blob. We propose a multi-view detection system that relies on a set of motionlet classifiers, which consist of detectors learned with vehicle samples clustered in the motion configuration space.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Attribute-based vehicle search in crowded surveillance videos

Feris

Siddiquie

Zhai

et al. 2011

Proceedings of the 1st ACM International Conference on Multimedia Retrieval

Self Cite

View full text Add to dashboard Cite

We present a novel application for searching for vehicles in surveillance videos based on semantic attributes. At the interface, the user specifies a set of vehicle characteristics (such as color, direction of travel, speed, length, height, etc.) and the system automatically retrieves video events that match the provided description. A key differentiating aspect of our system is the ability to handle challenging urban conditions such as high volumes of activity and environmental factors. This is achieved through a novel multi-view vehicle detection approach which relies on what we call motionlet classifiers, i.e. classifiers that are learned with vehicle samples clustered in the motion configuration space. We employ massively parallel feature selection to learn compact and accurate motionlet detectors. Moreover, in order to deal with different vehicle types (buses, trucks, SUVs, cars), we learn the motionlet detectors in a shape-free appearance space, where all training samples are resized to the same aspect ratio, and then during test time the aspect ratio of the sliding window is changed to allow the detection of different vehicle types. Once a vehicle is detected and tracked over the video, fine-grained attributes are extracted and ingested into a database to allow future search queries such as "Show me all blue trucks larger than 7ft length traveling at high speed northbound last Saturday, from 2pm to 5pm". See video demos of our system at:

show abstract

Section: Attribute Extraction / Searchmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Attribute-based vehicle search in crowded surveillance videos

Feris

Siddiquie

Zhai

et al. 2011

Proceedings of the 1st ACM International Conference on Multimedia Retrieval

Self Cite

View full text Add to dashboard Cite

show abstract

“…As an integrated part of the authors' recent line of research on view invariant multisite situation awareness, the paper presents a novel approach to solving the challenging problem of tracking multi-person across a multi-camera site, which is based on the "detect and track paradigm" while capitalising on the properties of auto-calibrated cameras as well as effective and efficient boosting classifiers. Lipton [20] report their work on "Searching surveillance video." The videos in the case of post-events forensic analysis will often come from a large number of geographically distributed cameras.…”

Section: The Industry Special Sessionmentioning

confidence: 99%

Issues in video analytics and surveillance systems: Research / prototyping vs. applications / user requirements

2007

2007 IEEE Conference on Advanced Video and Signal Based Surveillance

View full text Add to dashboard Cite

The next generation video surveillance system will be a networked, intelligent, multi-camera co-operative system with integrated situation awareness of dynamic and complex scenes, be it a wide area like an urban centre or indoor complex like a shopping mall. The core of such a system is the increasingly intelligent and robust video analytics that are capable of analysing the videos from low-level image appearance and feature extraction to midlevel object / event detection to high-level reasoning and scene understanding. While great strides have been made in researching the issues underlying the above capabilities by academia and industry labs over the last decade, the race is now on the applications of these in integrated systems, providing automated solutions for realistic surveillance problems. This panel discussion paper attempts to outline some of the issues arising from such a transformation, the way forward towards meeting end users' needs, and the aspects of on-going industry R&D activities. Questions are raised as to how to bridge the gap between the video analytics research and system applications and solutions in the market place.

show abstract

“…With a limited number of security personnel to watch vast amounts of video footage, law enforcement agencies have been increasingly relying on video analytics to automatically detect events of interest and spot abnormal behavior [1].…”

Section: Introductionmentioning

confidence: 99%

Appearance-Based Object Detection Under Varying Environmental Conditions

Feris

Brown

Pankanti

et al. 2014

2014 22nd International Conference on Pattern Recognition

Self Cite

View full text Add to dashboard Cite

Practical surveillance systems deployed in urban scenarios need to operate 24/7 under a wide range of environmental conditions. As modern video analytics shift from blob-based to object-centered architectures, appearance-based object detection under different weather conditions and lighting effects emerges as a critical yet largely unaddressed problem. This paper investigates this research topic, using as a case study the problem of vehicle detection in urban surveillance environments. In particular, we show that a simple and efficient Winsorized lighting correction technique improves performance significantly when outliers due to shadows, specularities, headlights, and occluders are present. Moreover, we demonstrate that a self-training mechanism utilizing a balanced training set automatically acquired from the target domain yields superior performance. Our experimental results are carried out on a novel dataset of vehicle images collected from a public traffic camera and categorized according to multiple environmental conditions.

show abstract

Searching surveillance video

Cited by 21 publications

References 8 publications

Attribute-based vehicle search in crowded surveillance videos

Attribute-based vehicle search in crowded surveillance videos

Issues in video analytics and surveillance systems: Research / prototyping vs. applications / user requirements

Appearance-Based Object Detection Under Varying Environmental Conditions

Contact Info

Product

Resources

About