Server-Based Person Detection Using Multi-Vehicle Image Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing techniques for detecting information on a person from an onboard camera require pre-registration of comparison data, limiting convenience as unregistered individuals cannot be detected.
Innovation Solution
An information processing system and method that involves vehicles generating and transmitting moving images of individuals outside the vehicle, with a server identifying and analyzing multiple images from different vehicles to detect behavior and estimate attribute information, improving detection accuracy and convenience.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If pre-registration of comparison data is required for person detection, then detection reliability is improved, but convenience and adaptability deteriorate as unregistered individuals cannot be detected
Solution Approach 1:
The system performs preliminary actions by collecting and storing moving images of persons in advance from multiple vehicles. These pre-collected images serve as a database that enables subsequent detection without requiring individual pre-registration, thus maintaining reliability while improving convenience.
Solution Approach 2:
The server system performs multiple functions: it collects images from various vehicles, stores them in a database, detects persons by comparing new images against the database, and estimates attributes. This multi-functional approach eliminates the need for pre-registration while maintaining detection reliability through universal image collection from multiple sources.
2Measurement precision
If multiple moving images from multiple vehicles are collected and analyzed, then detection accuracy and attribute estimation are improved, but system complexity and data processing requirements increase
Solution Approach 1:
The system merges moving images from multiple different vehicles into a unified database on the server. By combining data sources, the system improves detection accuracy through multiple viewing angles and more comprehensive person data, while the server centralizes processing to manage complexity.
Solution Approach 2:
The server acts as an intermediary that receives, stores, and processes images from multiple vehicles. It mediates between the complex data collection from various sources and the simpler detection task, managing the complexity by providing a centralized processing platform that handles image storage, comparison, and attribute estimation.
3Ease of operation
If behavior detection and attribute estimation are performed from multiple target moving images, then convenience for unregistered persons is improved, but information processing time and computational requirements increase
Solution Approach 1:
The system performs preliminary actions by collecting and storing moving images in advance, creating a pre-built database of person appearances and behaviors. When a new image is received, the system can quickly compare against this pre-collected data and estimate attributes without requiring time-consuming processing, thus reducing actual detection time while maintaining convenience for unregistered persons.
Data Source
AI summary
An information processing system includes a plurality of vehicle; and a server that is able to communicate with the plurality of vehicles. Each of the plurality of vehicles performs: generating a moving image in which a person outside the vehicle appears; and transmitting the moving image and an imaging time and an imaging position of the moving image to the server. The server performs: specifying two or more target moving images in which the same person appears out of a plurality of moving images received from the plurality of vehicles; detecting behavior of the person from the two or more target moving images; estimating attribute information of the person based on the detected behavior; and transmitting the estimated attribute information of the person and the imaging time and the imaging position of at least one of the target moving images to a client.