Video Content Filtering via Temporal Track Activity Scoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video surveillance systems require manual evaluation of video sequences to identify specific behaviors, which is cumbersome and inefficient, especially when longer tracks of actors are involved, as they demand extensive manual labeling of diverse and unusual behavior expressions.
Innovation Solution
A method for filtering input video sequences by detecting temporal tracks, assigning activity scores, computing event scores based on predefined activity types, and selectively generating output sequences that highlight relevant segments, allowing for automated selection and highlighting of specific events.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual evaluation of video sequences is used to identify specific behaviors, then detection accuracy can be maintained, but the evaluation process becomes cumbersome and inefficient
Solution Approach 1:
The system enables automated self-evaluation of video sequences by computing event scores based on temporal sequences of activity types. The processing system automatically selects relevant video segments without human intervention, using predefined score functions that evaluate track segments and compute compound scores for complex events, thereby resolving the contradiction between maintaining detection accuracy and improving evaluation efficiency
Solution Approach 2:
The patent replaces the mechanical manual evaluation process with an automated computational system. The processing system uses algorithmic score computations, temporal sequence analysis, and automated segment selection to substitute human operators, achieving both high detection accuracy through sophisticated scoring mechanisms and improved productivity through automation
2Reliability
If longer tracks of actors are used for behavior detection, then more comprehensive behavior analysis is achieved, but extensive manual labeling of diverse behavior expressions is required
Solution Approach 1:
The system segments long actor tracks into smaller track segments and applies localized activity score functions to each segment. This segmentation approach allows comprehensive analysis of long-duration behaviors by breaking them into manageable parts, each evaluated by appropriate score functions, thereby achieving reliable behavior analysis without requiring manual labeling of entire long sequences
Solution Approach 2:
The patent employs different activity score functions with varying parameters for different activity types and track segment characteristics. By dynamically adjusting scoring parameters based on the specific segment and activity type, the system achieves comprehensive behavior analysis from long tracks while avoiding the need for extensive manual labeling, as the parameterized functions automatically adapt to diverse behavior expressions
3Ease of operation
If automated selection of video segments is implemented, then the burden on human operators is reduced, but complex score computations are required
Solution Approach 1:
The system performs preliminary computations by pre-defining activity score functions and event score computations before actual video analysis. This preliminary setup includes establishing temporal sequences of activity types and configuring scoring parameters in advance, which simplifies the operational phase where segments are automatically selected based on pre-computed scores, thereby reducing operator burden while managing computation complexity through advance preparation
Solution Approach 2:
The patent introduces intermediate event scores as mediators between raw track data and final segment selection. These intermediate scores aggregate information from multiple activity types and temporal sequences, serving as a computational intermediary that simplifies the decision-making process for segment selection, reducing operator burden while structuring the complex computations in a manageable hierarchical manner
Data Source
AI summary
An input video sequence from a camera is filtered by a process that comprises detecting temporal tracks of moving image parts from the input video sequence and assigning activity scores to temporal segments of the tracks, using respective predefined track dependent activity score functions for a plurality of different activity types. Based on this, event scores for are computed as a function of time. This computation is controlled by a definition of a temporal sequence of activity types or compound activity types for an event type. Successive intermediate scores are computed, each as a function of time for a respective activity types or compound activity types in the temporal sequence. The successive intermediate scores for each respective activity types or compound activity are computed from a combination of the intermediate score for a preceding activity type or compound activity type in the temporal sequence at a preceding time and activity scores that were assigned to segments of the tracks after the preceding time, for the activity type or activity types defined by the compound activity type defined by the respective activity types or compound activity types in the temporal sequence. One of the computed event scores for a selected time. The computation of the selected event score is traced back to identify intermediate scores that were used to compute the selected one of the event scores and to identify segments of the tracks for which the assigned activity scores were used to compute the identified intermediate scores. An output video sequence and/or video image is generates that selectively includes the image parts associated with the selected segments.