GSFT-PSNR_Global_Single_Fuzzy_Threshold

spotify

PDF Download

GSFT-PSNR Global Single Fuzzy Threshold

Mind Map: GSFT-PSNR Global Single Fuzzy Threshold

1. Introduction

2. Key Contributions

3. Materials and Methods

3.1 One Level Thresholding

3.2 Multilevel Thresholding

3.3 Proposed Method (GSFT-PSNR)

4. PSNR Equation

5. Results and Discussion

6. Conclusion

7. Applications

GSFT-PSNR: Global Single Fuzzy Threshold Based on PSNR for OCR Systems

Abstract

Binarization or thresholding is a critical step in computer vision and image analysis, particularly in applications such as OCR (Optical Character Recognition) and augmented reality. This method separates the foreground from the background in an image, reducing the amount of data to process and improving computational efficiency. Traditional methods like Otsu’s thresholding are compared to a new method called GSFT-PSNR, which uses Peak Signal-to-Noise Ratio (PSNR) and fuzzy logic to determine threshold values. The proposed method shows improvements in various real-world applications such as license plate recognition and handwritten image processing.

Keywords

Introduction

Thresholding is a key first step in many computer vision applications, crucial for object recognition, camera calibration, and reducing noise. While single-level thresholding is fast and effective, multilevel thresholding can lead to higher computational time. Global thresholding methods are often favored for their speed, while local thresholding techniques are necessary for more complex scenes.

Key Contributions

The proposed method, GSFT-PSNR, combines global single fuzzy thresholding with PSNR to improve performance in varied lighting and environmental conditions. It is designed to adapt to different ambient illuminations and applications, such as OCR and license plate recognition. GSFT-PSNR excels in handling unstructured environments with variable lighting.

Materials and Methods

1. One Level Thresholding

2. Multilevel Thresholding

3. Proposed Method

PSNR Equation

[ PSNR = 10 \cdot \log_{10}\left(\frac{MAX^2}{MSE}\right) ]
Where MAX is the maximum pixel value, and MSE is the mean square error between the original and thresholded image.

Results and Discussion

Conclusion

The GSFT-PSNR method offers a robust, adaptive thresholding approach that can handle challenging real-world scenarios with varying lighting and complex scenes. By using PSNR as a quality measure and combining it with fuzzy logic, the method improves upon traditional thresholding techniques in OCR and other computer vision tasks.