Related Experiment Video
Updated: May 6, 2026
11:23
Lensless Fluorescent Microscopy on a Chip
Published on: August 17, 2011
17.8K
TSLEPS: A Two-Stage Localization and Erasure Method for Privacy Protection in Sensor-Captured Images
Xiaoxu Li1,2, Jun Fu3, Jinjian Wang1,2
1Key Laboratory of Tarim Oasis Agriculture, Ministry of Education, Tarim University, Alar 843300, China.
Sensors (Basel, Switzerland)
|August 28, 2025
Summary
This study introduces a Two-Stage Localization and Erasure method for Privacy protection in Sensor-captured images (TSLEPS). TSLEPS effectively detects and removes private text from images, preserving visual quality for mobile devices.
Area of Science:
- Computer Vision
- Image Processing
- Cybersecurity
Background:
- Increasing use of mobile imaging sensors raises concerns about image privacy leakage.
- Current manual privacy protection methods are inefficient and degrade image quality.
Purpose of the Study:
- To develop an automated and efficient method for protecting sensitive text information in sensor-captured images.
- To maintain image integrity and usability after privacy protection.
Main Methods:
- Proposed TSLEPS (Two-Stage Localization and Erasure method for Privacy protection in Sensor-captured images) with a two-stage framework.
- Utilized an inverted residual attention mechanism and generalized efficient aggregation layer for target detection.
- Employed a texture-enhanced feature attention mechanism and adversarial generative network for text erasure.
- Incorporated half-instance normalization for computational efficiency on mobile devices.
Main Results:
- Achieved 97.5% accuracy and 96.4% recall in privacy target detection.
- Reached 38.2140 dB PSNR and 0.9607 SSIM for privacy text erasure quality.
- Demonstrated effective performance on public real-world privacy datasets.
Conclusions:
- TSLEPS offers an effective solution for privacy protection in sensor-captured images.
- The method significantly improves detection accuracy, erasure quality, and computational efficiency.
- TSLEPS is suitable for deployment on resource-constrained mobile devices.
