Image Processing for Mouse Identification With 3D Skeleton Tracking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional technologies fail to identify individual animals, especially mice, in situations where they are constantly moving or multiple animals are present, and require additional equipment like cameras and thermographs, increasing costs and limiting applications to health management.
Innovation Solution
An information processing device that uses a skeleton estimation model to extract animal parts from images, combined with individual identification models, to identify and analyze the activity of mice within a predetermined range, utilizing machine learning to track and determine their activities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional biometric processing technology is used to measure animal data contactlessly, then measurement capability is achieved, but individual identification of moving animals or multiple animals cannot be performed
Solution Approach 1:
The patent segments the animal's body into multiple parts (head, body,四肢) and tracks each part independently through skeleton estimation. This segmentation allows the system to maintain identification reliability even when animals are moving or multiple animals are present, as each body part provides independent tracking information that can be combined to confirm individual identity.
Solution Approach 2:
The patent transitions from two-dimensional image analysis to three-dimensional skeleton representation. By estimating the spatial coordinates (x, y, z) of body parts and creating three-dimensional skeleton models, the system achieves reliable individual identification in three-dimensional space, enabling accurate tracking of moving animals and distinction between multiple animals overlapping in 2D images.
2Measurement precision
If a camera and thermograph are used to check health condition, then health monitoring capability is achieved, but device cost increases
Solution Approach 1:
The patent makes the skeleton estimation model multi-functional by enabling it to perform both individual identification and activity determination. The same three-dimensional skeleton data used for identifying animals also serves to analyze their activities (walking, running, jumping, grooming, etc.). This universality eliminates the need for separate health monitoring equipment like thermographs, reducing device cost while maintaining measurement precision for both identification and activity analysis.
Solution Approach 2:
The system uses the captured images themselves to derive multiple types of information. By processing the image data through skeleton estimation, the system simultaneously obtains individual identification information and activity information without requiring additional sensors or equipment. The image data serves multiple purposes, making the system self-sufficient and cost-effective.
3Ease of manufacture
If conventional technology is used for animal identification, then simple setup is achieved, but application scope is limited to health management only
Solution Approach 1:
The patent introduces dynamic activity determination that adapts to different animal behaviors. The system continuously monitors the three-dimensional skeleton data and dynamically determines activities such as walking, running, jumping, grooming, and eating based on movement patterns. This dynamic capability expands the application scope from static health monitoring to active behavior analysis, while maintaining simple setup requirements.
Solution Approach 2:
The patent changes the analysis parameters from static biometric measurements to dynamic spatial-temporal skeleton parameters. By tracking the positions and movements of body parts over time in three-dimensional space, the system can determine various activities and behaviors. This parameter transformation expands applicability to activity analysis, sociality studies, and behavioral research while keeping the system setup simple.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A challenge is to provide an art to identify mice as individuals from captured images of one or more mice that are doing activity in a predetermined activity range. An image processing device 2 includes an image acquisition unit 51, a part extraction unit 52, and an individual identification unit 54. The image acquisition unit 51 acquires a captured video of a state in which one or more animals are doing activity in a predetermined activity range. The part extraction unit 52 extracts multiple parts of the body of each of the one or more animals from each of unit images included in the video. The individual identification unit 54 analyzes the parts extracted from the unit images by the part extraction unit 52 in a time-series manner and identifies the one or more animals as individuals in each of the unit images on the basis of the analysis results. Thus, the challenge is solved.