De-interlacing algorithm responsive to edge pattern

• US 7,268,822 B2
• Filed: 07/23/2004
• Issued: 09/11/2007
• Est. Priority Date: 08/12/2003
• Status: Active Grant

• Current Assignee: Samsung Electronics Co. Ltd.
• Original Assignee: Samsung Electronics Co. Ltd.
• Inventors: Park, Min Kyu, Kang, Moon Gi
• Primary Examiner(s): Natnael; Paulos M.
• Application Number:
  US10/898,526
• Publication Number:
  US 20050036062A1
• Time in Patent Office: 1,145 Days
• Field of Search: 348/448, 348/441, 348/449, 348/451, 348/452, 348/458, 348/459, 375/240.21
• US Class Current: 348/448
• CPC Class Codes: H04N 5/142 Edging; Contouring
• CPC Class Codes: H04N 5/21 Circuitry for suppressing or minimising disturbance, e.g. moiré or halo
• CPC Class Codes: H04N 7/012 Conversion between an interlaced and a progressive signal for graphics images G09G2310/0229
• CPC Class Codes: H04N 7/0135 involving interpolation processes interpolation-based image scaling G06T3/4007; interpolation for video coding H04N19/587, H04N19/59
• CPC Class Codes: H04N 7/0142 the interpolation being edge adaptive edge detection in television signals H04N5/142; edge-driven scaling G06T3/403
• CPC Class Codes: H04N 7/0145 the interpolation being class adaptive, i.e. it uses the information of class which is determined for a pixel based upon certain characteristics of the neighbouring pixels

Abstract

There is provided a de-interlacing method including: receiving first and third image lines; selecting an upper vector having an N number of pixels among the first image lines, and selecting a lower vector having an N number of pixels among the third image lines; obtaining a weighted value on the basis of a relation between the pixels within the upper vector and a relation between the pixels within the lower vector; selecting an edge direction from the selected upper vector, the selected lower vector, and the obtained weighted value; and interpolating a pixel belonging to a second image line between the first and third image lines on the basis of the selected edge direction.

Claims

1. A de-interlacing method comprising:

receiving first and third image lines;

selecting an upper vector having an N number of pixels among the first image lines, and selecting a lower vector having an N number of pixels among the third image lines;

obtaining a weighted value on the basis of a relation between the pixels within the upper vector and a relation between the pixels within the lower vector;

selecting an edge direction from the selected upper vector, the selected lower vector, and the obtained weighted value; and

interpolating a pixel belonging to a second image line between the first and third image lines on the basis of the selected edge direction,

wherein the weighted value obtaining step comprises;

classifying an upper image pattern from the relation between the pixels within the upper vector, and classifying a lower image pattern from the relation between the pixels within the lower vector; and

obtaining the weighted value from the classified upper image pattern and lower image pattern.

2. The de-interlacing method of claim 1, wherein the upper image pattern is classified depending on a brightness variation between adjacent pixels within the upper vector, and the lower image pattern is classified depending on a brightness variation between adjacent pixels within the lower vector.

3. The de-interlacing method of claim 1, wherein the upper image pattern is classified depending on a comparison result of a threshold value and a brightness difference between the adjacent pixels within the upper vector, and the lower image pattern is classified depending on a comparison result of a threshold value and a brightness difference between the adjacent pixels within the lower vector.

4. The de-interlacing method of claim 3, wherein the upper image pattern has (N−

1) number of symbols respectively defined as increase (I), decrease (D) or fixity (F) depending on a brightness variation between two adjacent pixels within the upper vector, and

wherein the lower image pattern has (N−

1) number of symbols respectively defined as increase (I), decrease (D) or fixity (F) depending on a brightness variation between two adjacent pixels within the lower vector.

5. The de-interlacing method of claim 4, wherein the weighted value is provided in (N−

1) number respectively corresponding to the (N−

1) number of symbols for the upper image pattern and the (N−

1) number of symbols for the lower image pattern.

6. The de-interlacing method of claim 5, wherein if an n (n=1, 2, . . . , N−

1)th symbol for the upper image pattern is the increase (I) and an nth symbol for the lower image pattern is the increase (I), an nth weighted value is set as a first value.

7. The de-interlacing method of claim 6, wherein if the n (n=1, 2, . . . , N−

1)th symbol for the upper image pattern is the decrease (D) and the nth symbol for the lower image pattern is the decrease (D), the nth weighted value is set as the first value.

8. The de-interlacing method of claim 7, wherein if the n (n=1, 2, . . . , N−

1)th symbol for the upper image pattern is a fixity (F) and the nth symbol for the lower image pattern is the increase (I) or the decrease (D), the nth weighted value is set as a second value.

9. The de-interlacing method of claim 8, wherein if the n (n=1, 2, . . . , N−

1)th symbol for the lower image pattern is the fixity (F) and the nth symbol for the upper image pattern is the increase (I) or the decrease (D), the nth weighted value is set as the second value.

10. The de-interlacing method of claim 9, wherein if the n (n=1, 2, . . . , N−

1)th symbol for the upper image pattern is the increase (I) and the nth symbol for the lower image pattern is the decrease (D), the nth weighted value is set as a third value.

11. The de-interlacing method of claim 10, wherein if the n (n=1, 2, . . . , N−

1)th symbol for the upper image pattern is the decrease (D) and the nth symbol for the lower image pattern is the increase (I), the nth weighted value is set as the third value.

12. The de-interlacing method of claim 11, wherein if the n (n=1, 2, . . . , N−

1)th symbol for the upper image pattern is the fixity (F) and the nth symbol for the lower image pattern is the fixity (F), the nth weighted value is set as the second value.

13. The de-interlacing method of claim 12, wherein the third value is larger than the second value, and the second value is larger than the first value.

14. The de-interlacing method of claim 13, wherein the edge direction selecting step comprises the step of multiplying the weighted values with an absolute value of a difference between the upper vector and the lower vector to calculate the difference between the upper vector and the lower vector.

15. The de-interlacing method of claim 14, further comprising the step of selecting the edge direction from the upper vector and the lower vector having a minimal difference between the upper vector and the lower vector.

16. The de-interlacing method of claim 3, wherein the threshold value is set to be proportional to a variance value for a local region of image data.

17. The de-interlacing method of claim 16, wherein the threshold value is proportional to a difference between brightnesses respectively corresponding to an upper 20% and a lower 20% of a histogram for the local region.

18. A de-interlacing device comprising:

first and second line memories for respectively storing first and third image lines therein;

an image pattern classifier for selecting an upper vector having N number of pixels among the first image line stored in the first line memory and classifying an upper image pattern from a relation between pixels within the selected upper vector, and selecting a lower vector having N number of pixels among the third image line stored in the second line memory and classifying a lower image pattern from a relation between pixels within the selected lower vector;

a weighted value determiner for obtaining at least one weighted value on the basis of the upper image pattern and the lower image pattern classified by the image pattern classifier;

an edge direction selector for selecting an edge direction from the selected upper vector, the selected lower vector and the obtained weighted value; and

an edge interpolator for interpolating a pixel belonging to a second image line between the first image line and the third image line on the basis of the selected edge direction.

19. The de-interlacing device of claim 18, wherein the image pattern classifier classifies the upper image pattern depending on a comparison result of a threshold value and a brightness difference between adjacent pixels within the upper vector, and classifies the lower image pattern depending on a comparison result of the threshold value and a brightness difference between adjacent pixels within the lower vector.

20. The de-interlacing device of claim 19, wherein the threshold value is set to be proportional to a variance value for a local region of image data.

21. The de-interlacing device of claim 20, wherein the threshold value is proportional to a difference between brightnesses respectively corresponding to an upper 20% and a lower 20% of a histogram for the local region.

22. A program storage device readable by machine, tangibly embodying a program of instructions executable by the machine to perform method steps for de-interlacing an image, the method steps comprising:

receiving first and third lines of the interlaced image;

selecting an upper vector having a plurality of pixels from the first image line, and selecting a lower vector having a plurality of pixels from the third image line;

obtaining a weighted value responsive to a relation between the pixels within the upper vector and a relation between the pixels within the lower vector;

selecting an edge direction responsive to the selected upper vector, the selected lower vector, and the obtained weighted value; and

interpolating a pixel for a second image line disposed between the first and third image lines that is responsive to the selected edge direction;

wherein the weighted value obtaining step comprises;

classifying an upper image pattern from the relation between the pixels within the upper vector, and classifying a lower image pattern from the relation between the pixels within the lower vector; and

obtaining the weighted value from the classified upper image pattern and lower image pattern.