Multi-Camera Search Processing for Capture Count Tracking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional techniques struggle to determine how many times a search target is captured by each camera when multiple cameras capture the same individual, making it difficult to accurately identify and display relevant search results.
Innovation Solution
A search processing device that tracks moving objects across multiple cameras, calculates similarity degrees, and provides display modes to easily identify cameras capturing the target and the number of captures, using a tracking module, calculation module, and display control module to optimize feature extraction and display information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If conventional image-based search techniques are used to display people resembling the search target, then search results can be displayed sequentially, but it becomes difficult to determine how many times the search target has been captured by each camera
Solution Approach 1:
The patent segments the search results display into two distinct modes: a first display mode that shows only people meeting a similarity threshold, and a second display mode that organizes results by camera with capture counts. This segmentation allows users to access capture frequency information without complicating the primary search function.
Solution Approach 2:
The patent implements dynamic switching between display modes based on user needs. The system can transition from a simple similarity-based display to a detailed camera-based display showing capture frequencies, providing flexibility in how information is presented without requiring a completely different system.
2Quantity of substance
If search results from multiple cameras are displayed together, then all captured instances are visible, but it becomes difficult to identify which cameras captured the target and how many times
Solution Approach 1:
The patent segments the display by camera source, organizing search results into groups corresponding to different cameras. Each group shows the capture count for that specific camera, making it easy to identify which cameras captured the target and how many times each captured it.
Solution Approach 2:
The patent adds a new dimension to the display organization by introducing camera identification as a grouping criterion. Instead of simply listing results by similarity, the system adds camera source as an organizational dimension, allowing users to easily see which cameras captured the target.
Data Source
AI summary
A search processing device includes a memory and one or more processors coupled to the memory. The one or more processors are configured to: detect and track moving objects contained in videos by associating the moving objects with each other in a time sequence; receive setting of a search target contained in the videos; calculate a first descriptor and second descriptors, and calculate first degrees of similarity between the first descriptor and the second descriptors, the first descriptor indicating a feature of the search target and the second descriptors indicating features of the moving objects; and receive selection between a first display mode and a second display mode, and display search results corresponding to the search target on a display unit in the selected display mode.