Person Recognition in Video Streams Reducing False Alerts
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video surveillance systems face challenges in accurately identifying and categorizing meaningful segments of video streams, often producing unnecessary alerts due to normal activities or trivial movements, making it difficult for users to distinguish important events.
Innovation Solution
A method for recognizing persons in video streams, involving detection, personally identifiable information analysis, and classification of individuals as known or unknown, with features like facial signature generation and pose identification, to provide timely and relevant notifications.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If motion detection sensitivity is set high to detect all movements, then detection coverage is improved, but false alerts from trivial movements (tree leaves, sunlight shifts) increase
Solution Approach 1:
The patent segments the detection process into multiple stages: initial motion detection triggers a review process where captured images are analyzed by reviewers to determine if they represent true events or trivial movements. This segmentation allows the system to maintain high sensitivity while filtering false alerts through hierarchical processing.
Solution Approach 2:
The patent introduces an intermediary review process between motion detection and alert generation. Captured images are reviewed by human reviewers or automated analysis systems that act as intermediaries to distinguish between significant events and trivial movements, thereby reducing false alerts while maintaining comprehensive detection coverage.
2Object-generated harmful factors
If motion detection sensitivity is set low to reduce false alerts, then false alerts are reduced, but important events may be missed
Solution Approach 1:
The patent implements dynamic sensitivity adjustment where the detection threshold is not fixed but adapts based on environmental context, time of day, and learned patterns of normal versus abnormal activities. This allows the system to maintain high detection accuracy for important events while dynamically reducing sensitivity to trivial movements in appropriate contexts.
Solution Approach 2:
The system changes detection parameters dynamically based on environmental conditions, historical data, and contextual information. By adjusting sensitivity parameters according to time of day, location, and learned behavioral patterns, the system maintains high event detection accuracy while minimizing false alerts from trivial movements.
3Reliability
If all motion segments are recorded for review to ensure no events are missed, then event detection completeness is improved, but user time and effort increase
Solution Approach 1:
The patent applies partial action by recording and presenting only a subset of motion segments for review - specifically those that meet certain criteria or are flagged by preliminary analysis. Instead of requiring users to review all motion segments, the system selectively presents relevant segments, reducing user review time while maintaining complete event detection through automated pre-screening.
Solution Approach 2:
The system performs preliminary automated analysis of motion segments before presenting them to users. This preliminary action includes detecting motion patterns, filtering trivial movements, and prioritizing segments based on their likelihood of representing significant events, thereby reducing the volume of content users must review while ensuring no important events are missed.
4Measurement precision
If detailed analysis of all video segments is performed to accurately identify events, then event identification accuracy is improved, but processing time and computational resources increase
Solution Approach 1:
The patent segments the video analysis process into multiple stages with increasing detail: initial motion detection, preliminary pattern recognition, and detailed analysis only for segments that pass earlier filters. This segmentation allows accurate identification of events while reducing overall processing time by applying detailed analysis only where necessary.
Solution Approach 2:
The system applies detailed analysis partially - only to video segments that are flagged as potentially significant by preliminary automated detection. By performing comprehensive analysis only on a subset of segments rather than all video data, the system maintains high event identification accuracy while significantly reducing total processing time and computational resource consumption.
Data Source
AI summary
A method for recognizing persons in video streams includes obtaining a live video stream, detecting a first person in the live video stream, determining from analysis of the live video stream first information that identifies an attribute of the first person, determining based on at least some of the first information that the first person is not identifiable to the computing system, storing at least some of the first information, receiving a user classification of the first person as being a stranger, and deleting the stored first information.