Object State Estimation via Merged Count and Flow Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing techniques for estimating the state of objects in a space, such as people counting and flow analysis, either fail to accurately determine the number of objects or the flow, leading to incomplete understanding of the space's state, particularly in crowded areas.
Innovation Solution
An information processing apparatus that includes a first unit for estimating the number of objects in set regions and a second unit for estimating the flow of objects, using a combination of deep neural networks for object number estimation and LSTM for flow estimation, integrating these results to accurately determine the state and detect abnormalities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a method using neural network is used to count the number of objects in an image, then the number of objects can be estimated, but the flow of objects cannot be understood
Solution Approach 1:
The patent combines object number estimation results with optical flow analysis into a unified state estimation system. The state estimation unit integrates both the count of objects and their motion characteristics to comprehensively determine the state of the object group, thereby simultaneously obtaining both number information and flow information.
2Loss of information
If a method using optical flow analysis is used to determine the non-steady state of a crowd, then the flow of objects can be understood, but the number of objects cannot be obtained at the same time
Solution Approach 1:
The system merges optical flow analysis with object counting by having the state estimation unit receive both the number of objects from the neural network estimation and the flow characteristics from optical flow analysis, then integrate these to determine the overall state including both density and motion patterns.
3Measurement precision
If only object number estimation is performed, then the number of objects can be counted, but the state of objects in the space cannot be fully understood
Solution Approach 1:
The state estimation unit combines object number estimation with optical flow analysis to create a comprehensive state understanding. This integration allows the system to determine both the quantity and the motion state of objects, enabling full characterization of the space state including congestion detection and flow pattern recognition.
4Loss of information
If only flow analysis is performed, then the flow of objects can be analyzed, but the number of objects cannot be obtained
Solution Approach 1:
The system merges object counting and flow analysis by having the state estimation unit integrate results from both the neural network-based object number estimation and the optical flow-based motion analysis, thereby simultaneously obtaining accurate object counts and comprehensive flow characteristics.
Data Source
AI summary
An information processing apparatus includes a first estimation unit configured to estimate, for each of a plurality of images successive in time series, the number of objects existing in each of a plurality of set regions, and a second estimation unit configured to estimate a flow of the objects existing in each of the plurality of regions based on a result of the estimation for each of the plurality of images by the first estimation unit.