18403893. METHOD AND APPARATUS FOR RECOGNIZING USER BASED ON ON-DEVICE TRAINING simplified abstract (SAMSUNG ELECTRONICS CO., LTD.)

METHOD AND APPARATUS FOR RECOGNIZING USER BASED ON ON-DEVICE TRAINING

Organization Name

SAMSUNG ELECTRONICS CO., LTD.

Inventor(s)

Dohwan Lee of Suwon-si (KR)

Kyuhong Kim of Seoul (KR)

Jaejoon Han of Seoul (KR)

METHOD AND APPARATUS FOR RECOGNIZING USER BASED ON ON-DEVICE TRAINING - A simplified explanation of the abstract

This abstract first appeared for US patent application 18403893 titled 'METHOD AND APPARATUS FOR RECOGNIZING USER BASED ON ON-DEVICE TRAINING

Simplified Explanation

The patent application describes an on-device training-based user recognition method that involves training a feature extractor using reference data related to generalized users and user data, then using this trained feature extractor to determine registration and test feature vectors for user recognition.

• On-device training-based user recognition method:

   * Involves training a feature extractor on a device.
   * Uses reference data corresponding to generalized users and user data for training.
   * Determines registration feature vector based on user input data.
   * Determines test feature vector based on test data input.
   * Performs user recognition based on comparing registration and test feature vectors.

Potential Applications

The technology can be applied in:

• Mobile device security
• Personalized user experiences
• Access control systems

Problems Solved

This technology addresses:

• User authentication challenges
• Personalized device settings
• Security vulnerabilities in user recognition systems

Benefits

The benefits of this technology include:

• Improved user recognition accuracy
• Enhanced security measures
• Customized user experiences

Potential Commercial Applications

This technology can be utilized in:

• Mobile device industry
• Biometric security solutions
• Smart home technology

Possible Prior Art

One possible prior art for this technology could be:

• Existing user recognition systems using machine learning algorithms.

How does this technology handle noisy input data?

The patent application does not specify how the method deals with noisy or inaccurate input data. This could be crucial in real-world scenarios where data may not always be perfect.

What is the computational overhead of implementing this method on a device?

The application does not mention the computational resources required to perform on-device training and user recognition. Understanding the computational demands of this technology is essential for practical implementation.

Original Abstract Submitted

An on-device training-based user recognition method includes performing on-device training on a feature extractor based on reference data corresponding to generalized users and user data, determining a registration feature vector based on an output from the feature extractor in response to the input of the user data, determining a test feature vector based on an output from the feature extractor in response to an input of test data, and performing user recognition on a test user based on a result of comparing the registration feature vector to the test feature vector.