A data mining technique for discovering distinct patterns of hand signs: implications in user training and computer...

Area of Science:

• Human-Computer Interaction
• Data Mining
• Computer Science

Background:

• Hand signs are a key input method for computers in specific applications.
• Effective recognition of hand signs requires user training and optimized computer interface design.
• Understanding sign distinguishability is crucial for both user training and interface development.

Purpose of the Study:

• To present a data mining technique for discovering distinct hand sign patterns from sensor data.
• To identify groups of hand signs that are indistinguishable by computers.
• To provide insights for improving user training and computer interface design for hand sign recognition.

Main Methods:

• Utilized data mining techniques to analyze sensor data of hand signs.
• Identified distinct patterns within the hand sign data.
• Derived groups of signs that are difficult for computers to differentiate.

Main Results:

• Successfully discovered patterns in hand sign sensor data.
• Identified specific hand signs that are prone to misrecognition by computer systems.
• Provided a basis for categorizing signs by their computer-recognizability.

Conclusions:

• The data mining approach effectively identifies patterns and indistinguishable signs.
• This information is valuable for optimizing user training programs for hand sign input.
• Findings can guide the design of more intuitive and user-friendly computer interfaces for hand sign recognition.