Video Event Timeline for Motion Review and False Positive Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video surveillance systems face challenges in efficiently identifying and presenting meaningful motion events due to issues with motion detection sensitivity settings, leading to excessive recording of trivial movements or missed important events, and require improved methods for real-time and intuitive event detection and presentation.
Innovation Solution
Implementing a system that uses machine learning to tailor motion event categories to specific settings, incorporates false positive suppression, and allows for zone-based event detection and categorization, enabling real-time notifications and alerts, and providing user interfaces for event review and filtering.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If motion detection sensitivity is set high, then important events are detected, but false positives increase significantly
Solution Approach 1:
The patent segments motion detection into multiple categorical events (e.g., human motion, animal motion, vehicle motion, natural elements motion) with different sensitivity thresholds. This allows the system to detect important events while filtering out false positives by categorizing and selectively processing different types of motion separately.
Solution Approach 2:
The system dynamically adjusts sensitivity thresholds based on location-specific characteristics and event categories. Different zones have customized sensitivity levels, and the system adapts thresholds based on learned patterns of normal versus abnormal motion, resolving the contradiction between high detection accuracy and low false positive rates.
2Quantity of substance
If motion detection sensitivity is set low, then false positives are reduced, but important events are missed
Solution Approach 1:
By dividing motion detection into multiple categorized events with different sensitivity levels, the system maintains low overall false positive rates while ensuring important events in each category are detected with appropriate sensitivity. Each category can be optimized independently.
Solution Approach 2:
The system changes sensitivity parameters based on event category and location. Different motion categories have different detection thresholds, allowing the system to maintain high detection accuracy for important events while keeping false positives low through parameter optimization.
3Productivity
If conventional motion detection is used, then video segments are selected for review, but the volume of uninteresting video remains too high
Solution Approach 1:
The patent segments video data into categorized motion events (human, animal, vehicle, natural elements). This segmentation allows reviewers to focus on specific categories of interest and filter out uninteresting categories, dramatically reducing the volume of video that requires manual review while maintaining productivity.
Solution Approach 2:
The system extracts and separates meaningful motion events from the bulk video data by categorizing them. Reviewers can selectively review only relevant categories, taking out the essential information from the overwhelming volume of raw video data.
4Measurement precision
If manual scanning of all video is performed, then no events are missed, but time consumption increases significantly
Solution Approach 1:
By segmenting video into categorized motion events, the system allows reviewers to quickly scan category summaries rather than manually reviewing all video. This maintains detection completeness for selected categories while dramatically reducing review time.
Solution Approach 2:
The system performs preliminary categorization and filtering of video data before presentation to reviewers. Motion events are pre-processed, categorized, and organized, so reviewers receive ready-sorted information rather than raw video, eliminating the need for manual scanning while maintaining completeness.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method for video reproduction of a remotely captured video feed output by a video camera is described. The method comprises displaying a video monitoring user interface on a display of a client device located remotely from the video camera, the video monitoring user interface including: (i) a first region displaying a live and/or recorded video feed from the video camera, and (ii) a second region displaying an event timeline. The event timeline includes: (i) a plurality of equally spaced time indicators each indicating a specific time, and (ii) a current video feed indicator indicating a temporal position of the video feed displayed in the first region. The event timeline includes a past time corresponding to the recorded video feed from the video camera and a current time corresponding to the live video feed from the video camera. The current video feed indicator is movable relative to the equally spaced time indicators to facilitate a change in the temporal position of the video feed displayed in the first region. One or more event indicators corresponding to one or more events previously detected in the live video feed are displayed on the event timeline. In response to a user selection of the event timeline at a past time temporal position: a recorded video feed that was recorded at the selected past time temporal position is obtained; and the obtained recorded video feed is displayed in the first region of the video monitoring user interface. In response to a user selection of the event timeline at the current time temporal position, the live video feed is displayed in the first region of the video monitoring user interface.