Squeez'In: Private Authentication on Smartphones based on Squeezing Gestures

Document Title

Squeez’In: Private Authentication on Smartphones based on Squeezing Gestures

Document Information

• Domain: Human-Computer Interaction, smartphone touchscreen authentication technology
• Keywords: squeezing gestures, touch pressure, capacitive sensing, motion authentication, user behavioral traits, privacy protection, biometrics, user experience

Research Background and Issues

• Identified Problems or Challenges:

  • Current smartphone authentication technologies are divided into code-based methods (e.g., PIN passwords, pattern unlocking) and biometric methods (e.g., fingerprint, Face ID). Code-based methods are susceptible to shoulder-surfing attacks, have limited password space, and users often struggle to remember complex passwords. Biometric methods, due to their immutable nature, pose privacy and security risks (e.g., data replication).
  • To address these issues, researchers have proposed several motion- and behavior-based authentication techniques. However, these methods often lack concealment in practical applications or lead to privacy concerns due to continuous tracking.

• Importance of the Research:

  • Authentication operations are among the most frequent interactions between users and smartphones in daily life. Efficient, secure, and natural authentication methods are critical in the mobile computing era.
  • Providing a more privacy-preserving and natural authentication method, especially one based on user-specific behavioral traits that are difficult to replicate, is vital for enhancing smartphone security and user acceptance.

• Research Motivation and Related Work:

  • Existing studies suggest that authentication based on touch pressure and behavioral traits has potential. These methods are generally based on dynamic behavioral traits but have not fully explored the possibilities of concealed interactions.
  • Squeezing gestures represent a novel motion-based authentication approach with high concealment and a larger design space for passwords.

Solution

• Proposed Method or Solution: This paper introduces a smartphone authentication technology based on squeezing gestures—Squeez’In. Users authenticate through customized squeezing patterns (length, pressure, duration), combined with user-specific behavioral traits.

• Innovative Contributions:

  1. Natural Design: Systematic exploration of user-defined squeezing gesture characteristics and proposed design guidelines.
  2. Behavioral Trait Integration: Enhanced security and non-replicability by combining squeezing behaviors with user-specific traits.
  3. Efficient Implementation: Authentication performed using data from standard capacitive touchscreens without relying on additional hardware.

• Implementation Steps and Key Technologies:

  1. User Study Design:
     • Research on user-defined gestures identified the most commonly used features in squeezing gestures, including gesture length (2-7), touch pressure (light and heavy), and touch duration (long and short).
     • Verified users' ability to control actions during squeezing, resulting in recommended feature levels.
  2. System Implementation:
     • Pressure sensing implemented on existing capacitive touchscreen smartphones. Developed authentication models based on Support Vector Machines (SVM) and Gradient Boosting Decision Trees (GBDT), utilizing data augmentation techniques to address sample scarcity.
  3. Performance Optimization:
     • Improved authentication performance through parameter optimization and feature engineering.
  4. Experimental Validation:
     • Conducted multiple rounds of user-simulated authentication and long-term user stability studies to validate practicality, longevity, and user preferences.

Research Outcomes

• Specific Achievements:

  1. High-Accuracy Authentication:
     • Achieved an authentication accuracy rate of 99.3%, F1-Score of 0.93, False Acceptance Rate (FAR) of 0.4%, and False Rejection Rate (FRR) of 6.4% across data from 21 users.
     • Successfully distinguished between users even when two users input identical gestures, with FAR as low as 1%.
  2. Long-Term Stability:
     • In a 14-day user stability test, gesture memory retention and recognition accuracy remained stable (average recognition error rate of 7.1%).
  3. User Experience Advantages:
     • Compared to PIN and pattern passwords, Squeez’In achieved authentication speeds 23-14% faster and received high user approval for privacy, security, and social acceptability.

• Advantages Over Existing Solutions:

  • Compared to static biometrics (e.g., fingerprint, Face ID), Squeez’In allows dynamic modification of gestures to mitigate risks of data theft.
  • Compared to graphical or password-based authentication, it supports implicit authentication behaviors and is highly resistant to shoulder-surfing attacks.

• Experimental or Evaluation Results:

  • Users demonstrated flexibility and intuitiveness in designing gesture lengths, pressures, and durations, without confusion during long-term use.
  • Experiments showed that Squeez’In’s registration and authentication steps were more efficient than biometric methods, supporting fast authentication while enhancing privacy.

• Limitations and Future Directions:

  1. Current implementation relies on curved screen smartphones; general implementation for flat screens needs exploration.
  2. Challenges remain in sensing squeezing postures and other gesture features (e.g., finger identity).
  3. The experimental population was relatively young and limited in sample size; broader testing in diverse populations and real-world applications is required.

Future Directions: Further optimization of model performance and user comfort, integration with other interaction technologies to support more complex application scenarios, and expansion of testing to various device types and user groups.