Sign Language Video Collection With Attribute-Based Word Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems for collecting sign language video data are biased and do not allow for even collection from a diverse range of individuals due to differences in sign language movements based on personal attributes.
Innovation Solution
A system that determines registration candidate words for sign language registrants based on their attributes, such as gender, age, and physique, and collects corresponding video data, ensuring an unbiased and even distribution of data across various attributes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If sign language video data is collected based on user free will selection, then users can voluntarily participate in data collection, but the collected data becomes biased and uneven across different individuals
Solution Approach 1:
The system performs preliminary action by pre-determining registration candidate words based on user attributes before the actual data collection process. The determination unit analyzes user information (age, gender, physique) and pre-selects appropriate words for registration, ensuring that data collection starts with a balanced and representative word selection rather than allowing completely free will selection.
Solution Approach 2:
The system implements feedback by continuously monitoring registration status of different words across users and adjusting the determination of registration candidate words accordingly. The determination unit refers to registration status information to dynamically select words that need more registration data, creating a feedback loop that progressively balances the data collection distribution.
2Manufacturing precision
If registration candidate words are determined by the system based on user attributes, then data collection becomes unbiased and even across different individuals, but the system complexity increases
Solution Approach 1:
The system applies segmentation by dividing the word registration process into distinct segments: word determination segment (where the determination unit selects candidate words based on attributes), registration segment (where users actually register the selected words), and status monitoring segment (where the unit tracks registration progress). This segmentation allows complex functionality to be managed through modular, manageable components.
Solution Approach 2:
The determination unit acts as an intermediary between the user's attribute information and the word registration process. It mediates by translating user characteristics (age, gender, physique) into appropriate word selection criteria, thereby simplifying the overall system architecture while achieving uniform data collection.
3Quantity of substance
If the system collects sign language video data from many people, then the data set becomes more comprehensive and unbiased, but the time required for data collection increases
Solution Approach 1:
The system performs preliminary word selection based on user attributes before data collection begins. By pre-determining which words each user should register based on their characteristics, the system avoids time-consuming trial-and-error selection processes and ensures that users immediately contribute to the most needed data categories.
Solution Approach 2:
The system uses feedback from registration status monitoring to dynamically adjust word selection and collection priorities. By identifying which words have insufficient registration data across different user attributes, the system can focus collection efforts on high-priority words, thereby accelerating the achievement of comprehensive data sets.
Data Source
AI summary
A sign language video information collecting apparatus includes a registration word candidate determining unit that determines registration candidate words for which a sign language registrant is requested to register sign language and outputs the determined registration candidate words to a terminal device of the sign language registrant, a sign language video information obtaining unit that obtains sign language video information corresponding to a registration candidate word which is output from the terminal device of the sign language registrant in response to input of the registration candidate words, and a sign language video information storage unit 14 that stores the sign language video information corresponding to the registration candidate word correlated with the registration candidate word.