Method for image resolution enhancement

• US 20040184657A1
• Filed: 03/18/2003
• Published: 09/23/2004
• Est. Priority Date: 03/18/2003
• Status: Active Grant

• Current Assignee: Advanced & Wise Technology Corp.
• Original Assignee: Advanced & Wise Technology Corp.
• Inventors: Lin, Chin-Teng, Chen, Jia-Lin, Liang, Sheng-Fu, Pu, Her-Chang
• Granted Patent: US 7,187,811 B2
• Time in Patent Office: Days
• US Class Current: 382/159
• CPC Class Codes: G06T 3/4007 based on interpolation, e.g. bilinear interpolation image demosaicing G06T3/4015; edge-driven or edge-based scaling G06T3/403
• CPC Class Codes: G06T 3/4046 using neural networks

Abstract

The invention provides a method for image resolution enhancement, which utilizes a fuzzy analysis system simulating the human vision system and uses the neural network as a basis for digital image interpolation. After an original image is inputted in, the image analysis module will divide and classify the original image, and then each of the image being classified will be processed by either the bilinear interpolation or the neural network interpolation. Because the fuzzy analysis system is configured according to the human vision system, and because the neural network is a model obtained from learning real natural images, the vision effect of the image enlarged through the processing method of the invention is very close to the real natural image.

Claims

1. A method for image resolution enhancement, including the following steps:

providing an original image;

utilizing a fuzzy analysis system to analyze the original image and classifying the original image into the one with edge quality and the other one without edge quality;

calculating the edge angle of the area classified to be with edge quality and processing the area with neural network interpolation;

processing the area classified to be without edge quality with bilinear interpolation; and

obtaining a high-resolution digital image after completing the whole original image processing.

2. The method for image resolution enhancement as claimed in claim 1, wherein an image mask can be used to divide the original image before the fuzzy analysis system start the analysis so that the subsequent procedures of analysis can be carried out smoothly.

3. The method for image resolution enhancement as claimed in claim 1, wherein the image mask is composed of N*N pixels.

4. The method for image resolution enhancement as claimed in claim 1, wherein the fuzzy analysis system includes an image analysis module for dividing and analyzing the original image, and a classification module for classifying the original image according to the analytical result from the image analysis module.

5. The method for image resolution enhancement as claimed in claim 4, wherein the image analysis module includes:

an image edge retrieving module, which retrieves the edge area of the original image; and

an angle calculation unit, which calculates the angle of each pixel on the edge area by utilizing the edge area with orientation.

6. The method for image resolution enhancement as claimed in claim 1, wherein there are three input variables being based upon by the fuzzy analysis system to perform classification, including visibility, structure, and complexity.

7. The method for image resolution enhancement as claimed in claim 1, wherein during the classification performed by the fuzzy analysis system, if the output variable is larger than or equal to the preset threshold value, the neural network interpolation should be applied to process the original image;

on the other hand, if the output variable is smaller than the preset threshold value, the bilinear interpolation should be applied.

8. The method for image resolution enhancement as claimed in claim 7, wherein the threshold value can be determined according to the digital image and its condition required for processing.

9. The method for image resolution enhancement as claimed in claim 1, wherein in the step of calculating the edge angle, an angle calculation module is applied to calculate the angle of the edge area.

10. The method for image resolution enhancement as claimed in claim 1, wherein the neural network interpolation utilizes the supervised learning method to train the neural network and stores the pertinent parameters in a database after the neural network training is completed.

11. A method for image resolution enhancement, including the following steps:

providing an original image;

retrieving the edge portion of the original image;

calculating the angle of each pixel on the image edge area;

classifying the original image into the one using the neural network interpolation and the other one using the bilinear interpolation according to the position of image edge area and the angle being calculated;

processing the original image by applying the neural network interpolation and the bilinear interpolation; and

finishing the whole original image processing to obtain a high-resolution digital image.

12. The method for image resolution enhancement as claimed in claim 11, wherein the image mask can be used to divide the original image before the step of retrieving the image edge area is executed so that the subsequent procedures of analysis can be carried out smoothly.

13. The method for image resolution enhancement as claimed in claim 12, wherein the image mask is composed of N*N pixels.

14. The method for image resolution enhancement as claimed in claim 11, wherein in the step of calculating the angle, an angle calculation module is applied to calculate the angle of edge area.

15. The method for image resolution enhancement as claimed in claim 11, wherein the neural network interpolation utilizes the supervised learning method to train the neural network and stores the pertinent parameters in a database after the neural network training is completed.