Person Recognition Using Temporal Feature Sequences
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional video monitoring systems face challenges in accurately recognizing individuals in large public areas due to low camera resolution, high intra-person variance, and environmental factors like lighting, leading to errors in person recognition.
Innovation Solution
A monitoring device that uses a feature determination apparatus to create feature vectors for objects in video images, combined with a person recognition apparatus that includes an association apparatus to expand the search by considering surrounding objects, enhancing recognition robustness by incorporating features from associated objects like family members or items.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If low resolution cameras are used to reduce data load in large monitoring areas, then data transmission load is reduced, but person recognition accuracy deteriorates
Solution Approach 1:
The patent transitions from analyzing single-image features to analyzing temporal sequences of images. By incorporating time as an additional dimension and using recurrent neural networks to process image sequences, the system achieves higher recognition accuracy without requiring higher spatial resolution cameras, thus maintaining low data transmission loads.
Solution Approach 2:
The system changes the parameter of analysis from static single-frame features to dynamic temporal feature sequences. By extracting features across multiple time frames and combining them through RNN processing, the system compensates for low spatial resolution through temporal information accumulation, improving recognition accuracy while keeping camera resolution low.
2Device complexity
If conventional facial recognition methods are used in low resolution videos, then implementation is simple, but recognition reliability deteriorates due to errors
Solution Approach 1:
The patent introduces dynamic temporal processing into the recognition system. Instead of static single-frame analysis, the system processes sequences of images through recurrent neural networks that capture temporal dynamics of human movement and appearance changes, significantly improving reliability while maintaining manageable system complexity through efficient sequence processing.
Solution Approach 2:
The system implements continuous temporal analysis by processing overlapping image sequences with sliding windows. This continuous processing ensures that recognition decisions are based on accumulated evidence over time rather than isolated frames, improving reliability while the incremental processing approach keeps computational complexity manageable.
3Reliability
If high intra-person variance is accommodated by using multiple features, then recognition robustness improves, but system complexity increases
Solution Approach 1:
The patent merges multiple feature extraction approaches (spatial features from images and temporal features from sequences) into a unified recurrent neural network framework. This integration allows the system to handle intra-person variance by combining complementary information sources while maintaining a cohesive architecture that manages complexity through unified processing rather than separate independent systems.
Data Source
AI summary
The invention relates to a monitoring device (10) for recognizing persons in a monitoring region (2), the monitoring region (2) being video-monitored by means of at least one camera (6) and the camera (6) being designed to provide monitoring images (7) to the monitoring device (10) as video data, the monitoring device comprising: —a feature determination apparatus (13), the feature determination apparatus (13) being designed to determine a feature vector (19) for each object in at least one of the monitoring images (7); —a person recognition apparatus (16), the person recognition apparatus (16) being designed to detect in the monitoring images (7) a person to be recognized (11), on the basis of the determined feature vector and/or the determined feature vectors (19) of the feature determination apparatus (13) and/or a combined feature vector (18); —an association apparatus (14), the association apparatus (14) being designed to determine a feature vector (19) for each person to be recognized (11) and each associated environment object of the person to be recognized (11), the association apparatus (14) being designed to determine the combined feature vector (18) on the basis of the feature vector (19) of the person to be recognized (11) and the feature vector or the feature vectors (20) of the associated environment objects.