This paper offers an approach to detecting unspecified anomalies and events in unspecified data streams. The underlying technology is spectrum independent and also does not depend on correlated pixels or other data to achieve accurate event detection and extraction in highly robust environments. The technical approach presented here uses a network of simple building-block equations to predict the non-event data and thereby present subtle sub-streams to a detection model as potential events of interest. The prediction model is automatically created from sequential observations of the data stream. Once model construction is complete, it continues to evolve as new samples are taken. Each sample value that is sufficiently different from the model's predicted value is postulated as an event. A subsequent detection model uses a simple set of rules to improve the prediction model's ability to more outliers. We review the theory of the model and its application in two demonstrations: intruder detection in a robust video scene and voice detection in a noisy audio signal. These demonstrations are coupled to a notional concept of operation that emphasizes the spectrum-independence of this approach and its integration with other processing requirements such as target recognition and tracking. The primary benefit of this technique is its ability to process large data volumes for obscured or buried information within highly active environments. The fully automated nature of this technique helps mitigate manning shortfalls typically associated with sorting through large volumes of surveillance data using trained analysts. Our approach enables an organization to perform automated cueing for these analysts so that they spend less time examining data where nothing of interest exists. This maximizes the value of skilled personnel by using them to assess data with true potential. In this way, larger data volumes could be processed in a shorter period of time leading to a higher likelihood that important events and signals will be found, analyzed, and acted upon.
Finding events automatically in continuously sampled data streams via anomaly detection
01.01.2000
1092426 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Finding Events Automatically in Continuous Sampled Data Streams Via Anomaly Detection
British Library Conference Proceedings | 2000
|ANALYZING RECURRENT STREAMS OF DIGITAL DATA TO DETECT AN ANOMALY
Europäisches Patentamt | 2020
|Method and system for automatically finding parking space and automatically parking vehicle
Europäisches Patentamt | 2022
|SCALABLE AND DISTRIBUTED DETECTION OF ROAD ANOMALY EVENTS
Europäisches Patentamt | 2021
|NTIS | 1969
|