Action Recognition Through Time-Series Element Concatenation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies require predefined basic actions to recognize human actions, leading to inappropriate recognition when these actions are undefined.
Innovation Solution
An action recognizing apparatus and method that extracts action feature data, converts it into element data, and concatenates it to generate basic action data, allowing recognition of actions without predefined definitions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If basic actions are defined in advance, then action recognition can be performed using predefined categories, but actions cannot be recognized when basic actions are not defined
Solution Approach 1:
The system performs self-service by automatically generating basic action definitions from the action data itself through unsupervised learning. The information processing apparatus autonomously extracts action elements and generates basic action definitions without requiring external predefinition, enabling the system to adapt to any action type automatically.
Solution Approach 2:
The system changes the parameter of basic action definitions from fixed predefined values to dynamically generated values based on the action data. By using unsupervised learning to extract action elements and generate basic actions, the system transforms the rigid predefined parameter into a flexible data-driven parameter that adapts to different action types.
2Ease of manufacture
If predefined basic actions are used, then the recognition system is simpler to implement, but it cannot recognize actions outside the predefined set
Solution Approach 1:
The system generates its own basic action definitions from the input action data through unsupervised learning, eliminating the need for manual predefinition while maintaining implementation simplicity. This self-service approach ensures the system can recognize any action type present in the data with high reliability.
Solution Approach 2:
The system transitions from static predefined basic actions to dynamic basic actions generated automatically from the action data. This dynamic generation allows the system to adapt to different action types and contexts, improving recognition reliability while maintaining ease of implementation through automated processes.
Data Source
AI summary
An action recognizing apparatus 100 of the present disclosure includes: an extracting unit 121 that extracts action feature data representing a feature of an action in each predetermined time unit from time-series action data; a converting unit 122 that converts the action feature data of each predetermined time unit into action element data; and a concatenating unit 123 that generates, as basic action data, concatenated data obtained by concatenating the action element data on a basis of a time-series array of the action element data. Thereby, basic actions can be recognized from action data, and can be used for assisting decision making.