ML Model Arrangement for Motion Pattern Evaluation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems for evaluating motion patterns in image data sequences lack the capability to accurately identify motion phases and evaluation points, particularly in complex motion patterns, and do not provide effective feedback for users performing exercises.
Innovation Solution
A machine learning model arrangement comprising a first ML model for predicting key data elements and at least one second ML model for classifying motion phases and evaluation points, using different types of ML models such as convolutional neural networks, decision trees, or random forests, to analyze and evaluate motion patterns in image data sequences.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single ML model is used for motion pattern evaluation, then the system complexity is low, but the measurement precision and reliability of motion phase identification deteriorate
Solution Approach 1:
The patent divides the motion evaluation task into two separate ML models: a first model for predicting key data elements (landmark positions) and a second model for evaluating motion patterns and identifying phases. This segmentation allows each model to specialize in its specific function, improving overall measurement precision while keeping individual model complexity manageable.
Solution Approach 2:
The first ML model acts as an intermediary that processes raw image data and extracts key data elements, which then serve as input features for the second ML model. This intermediate processing step enables the second model to focus specifically on motion pattern recognition without being burdened by raw data processing, thereby improving reliability.
2Device complexity
If a single ML model is used for motion pattern evaluation, then the device complexity is low, but the reliability of evaluation deteriorates
Solution Approach 1:
The evaluation system is segmented into two specialized models: one for feature extraction (key data elements) and another for motion pattern recognition. This functional segmentation improves reliability by ensuring that each component is optimized for its specific task, reducing errors that would occur in a single general-purpose model.
Solution Approach 2:
The first ML model automatically extracts and prepares key data elements from image sequences, which then serve as refined input for the second model's motion evaluation. This self-service mechanism ensures that the evaluation process receives high-quality, pre-processed data, thereby improving overall system reliability.
3Measurement precision
If different types of ML models are used for different tasks, then the measurement precision improves, but the device complexity increases
Solution Approach 1:
The patent applies different types of ML models suited to specific local tasks: a first model type optimized for predicting spatial key data elements and a second model type optimized for temporal motion pattern classification. This local optimization of model types to match specific task requirements improves measurement precision while the modular architecture keeps the overall complexity manageable.
Data Source
AI summary
A ML model arrangement configured for evaluating motion patterns in a sequence of image data structures is described. The ML model arrangement comprises a first ML model configured for predicting a set of key data elements for each image data structure of the sequence of image data structures, a key data element indicating a respective position of a landmark in the image data structure. The ML model arrangement further comprises at least one second ML model, each second ML model being a ML model configured for evaluating a corresponding specific motion pattern. Each second ML model is configured for determining, based on input data comprising at least one of the key data elements predicted for at least one image data structure or data derived therefrom, class labels for each image data structure, said class labels identifying at least one of: at least one motion phase of the specific motion pattern, at least one evaluation point of the specific motion pattern.