Fast edge-oriented image interpolation algorithm

• US 20050058371A1
• Filed: 09/11/2003
• Published: 03/17/2005
• Est. Priority Date: 09/11/2003
• Status: Active Grant

• Current Assignee: Chevron Technology Ventures LLC (Chevron Corp.)
• Original Assignee: Chevron Technology Ventures LLC (Chevron Corp.)
• Inventors: Huang, Chin-Hui, Chen, Mei-Juan
• Granted Patent: US 7,286,721 B2
• Time in Patent Office: Days
• US Class Current: 382/300
• CPC Class Codes: G06T 3/403 Edge-driven scaling; Edge-based scaling

Abstract

The invention provides an image interpolating method, which comprises receiving the low resolution pixels Y_ij. Then, a homogenous area and an edge area of the image are determined according to pixel differences of the pixels Y_{2i, 2j} in comparing with a threshold. Then, the pixels Y_2i,2j belonging to the homogenous area are interpolated by a first interpolating algorithm, while the pixels Y_2i,2j belonging to the edge area are interpolated by a second interpolating algorithm.

Claims

1. An image interpolating method, wherein low resolution pixels Y_ij of an image are zoomed to high resolution pixels Y_2i,2j, comprising:

receiving the low resolution pixels Y_ij;

determining a homogenous area and an edge area of the image based on pixel differences of the pixels Y_2i,2j in comparing with a threshold;

interpolating the low resolution pixels belonging to the homogenous area into the high resolution pixels by a first interpolating algorithm; and

interpolating the low resolution pixels belonging to the edge area into the high resolution pixels by a second interpolating algorithm.

2. The image interpolating method of claim 1, in the step of determining the homogenous area and the edge area of the image, wherein three variables of

Δ

Y₁=|Y_2i,2j−

Y_2i+2p,2j+2q|, p,qε

{(0,1), (1,0)},
Δ

Y₂=|Y_2i+2,2j−

Y_2i,2j+2|, and
Δ

Y₃=|Y_2i,2j−

Y_2i+2,2j+2|

are used to determine whether the homogenous area or the edge area by a condition set of;

if Δ

Y₁<

the threshold then

the pixel Y_2i+p,2j+q is in the homogenous area

else

the pixel Y_2i+p,2j+q is in the edge area as one of edge pixels;

if Δ

Y₂<

the threshold and Δ

Y₃<

the threshold then

the pixel Y_2i+1,2j+1 is in the homogenous area

else if Δ

Y₂<

the threshold then

the pixel Y_2i+1,2j+1 is in the homogenous area

else if Δ

Y₃<

the threshold then

the pixel Y_2i+1,2j+1 is in the homogenous area

else

the pixel Y_2i+1,2j+1 is in the edge area as one of edge pixels.

3. The image interpolating method of claim 2, wherein the first interpolating algorithm includes obtaining the pixel Y_2i+p,2j+q by calculating (Y_2i,2j+Y_2i+2p,2j+2q)/2.

4. The image interpolating method of claim 2, wherein the first interpolating algorithm includes:

when Δ

Y₂<

the threshold and Δ

Y₃<

the threshold,

the pixel Y_2i+1,2j+1 is obtained by calculating Y_2i+1,2j+1=(Y_2i+2,2j+Y_2i,2j+2)/2 if the Δ

Y₂ less than Δ

Y₃; and

the pixel Y_2i+1,2j+1 is obtained by calculating Y_2i+1,2j+1=(Y_2i,2j+Y_2i+2,2j+2)/2 if the Δ

Y₃ is less than Δ

Y₂.

5. The image interpolating method of claim 2, wherein the first interpolating algorithm includes:

when only Δ

Y₂<

the threshold for the Δ

Y₂ and the Δ

Y₃, the pixel Y_2i+1,2j+1 is obtained by calculating (Y_2i+2,2j+Y_2i,2j+2)/2.

6. The image interpolating method of claim 2, wherein the first interpolating algorithm includes:

when only Δ

Y₃<

the threshold for the Δ

Y₂ and the Δ

Y₃, the pixel Y_2i+1,2j+1 is obtained by calculating (Y_2i,2j+Y_2i+2,2j+2)/2.

7. The image interpolating method of claim 1, wherein the first interpolating algorithm includes:

when the pixels Y_{2i, 2j} in the homogenous area, the pixels Y_{2i, 2j} are interpolated by a linear interpolation algorithm.

8. The image interpolating method of claim 2, wherein the second interpolating algorithm includes interpolating the pixels Y_2i,2j along a direction having a minimum difference in the neighboring pixels.

9. The image interpolating method of claim 8, wherein the neighboring pixels of one of the pixels Y_2i,2j does not include a determined edge pixel.

10. The image interpolating method of claim 8, wherein when the minimum difference diff_min is determined by taking a minimum of four differences of

diff₁=|Y_2i−

1,2j−

Y_2i+1,2j|,
diff₂=|Y<sub>2i−

1,2j−

1</sub>−

Y_2i+1,2j+1|,
diff₃=|Y<sub>2i,2j−

1</sub>−

Y_2i,2j+1|, and
diff₄=|Y<sub>2i+1,2j−

1</sub>−

Y<sub>2i−

1,2j+1</sub>|,

wherein the differences including one of the edge pixels is skipped.

11. The image interpolating method of claim 8, wherein the pixel Y_ij is obtained by calculating (Y_2i−

1,2j+Y_2i+1,2j)/2 at a direction with the minimum pixel difference.

13. The image interpolating algorithm of claim 1, further comprising interpolating the edge pixels according to a second algorithm as follows:

calculating a plurality of pixel differences of

diff₁=|Y<sub>2i−

1,2j</sub>−

Y_2i+1,2j|,
diff₂=|Y<sub>2i−

1,2j−

1</sub>−

Y_2i+1,2j+1|,
diff₃=|Y<sub>2i,2j−

1</sub>−

Y_2i,2j+1|, and
diff₄=|Y<sub>2i+1,2j−

1</sub>−

Y<sub>2i−

1,2j+1</sub>|,

wherein the differences including one of the edge pixels is skipped;

finding a minimum of the pixel differences; and

interpolating the pixel Y_{2i, 2j}=(Y<sub>2i−

1,2j</sub>+Y_2i+1,2j)/2 at a direction with the minimum pixel difference.

12. An image interpolating algorithm for an image, wherein low resolution pixels Y_{i, j} of the image are zoomed to high resolution pixels Y_{2i, 2j}, wherein three variables of Δ

Y₁=|Y_2i,2j−

Y_2i+2p,2j+2q|, Δ

Y₂=|Y_2i+2,2j−

Y_2i,2j+2|, and Δ

Y₃=|Y_2i,2j−

Y_2i+2,2j+2|, p,q ε

{(0,1), (1,0)} are used, the image interpolating algorithm comprising;

determining at least one of edge pixel and interpolating the pixels Y2i, 2j if the pixel to be interpolated is not the edge pixel by a first algorithm as follows;

if Δ

Y1<

a threshold then

Y_2i+p,2j+q=(Y_2i,2j+Y_2i+2p,2j+2q)/2

else

Y_2i+p,2j+q are the edge pixel

if Δ

Y₂<

the threshold and Δ

Y₃<

the threshold then

Δ

Y_min=min{Δ

Y₂, Δ

Y₃}

if Δ

Y_min=Δ

Y₂

Y_2i+1,2j+1=(Y_2i+2,2j+Y_2i,2j+2)/2

else

Y_2i+1,2j+1=(Y_2i,2j+Y_2i+2,2j+2)/2

else if Δ

Y₂<

the threshold then

Y_2i+1,2j+1=(Y_2i+2,2j+Y_2i,2j+2)/2

else if Δ

Y₃<

the threshold then

Y_2i+1,2j+1=(Y_2i,2j+Y_2i+2,2j+2)/2

else

Y_2i+1,2j+1 is one of the edge pixel.