A Zero-Value Prediction Technique for Fast DCT Computation

1 / 6

# A Zero-Value Prediction Technique for Fast DCT Computation - PowerPoint PPT Presentation

A Zero-Value Prediction Technique for Fast DCT Computation. Y. Nishida, K. Inoue, and V. Moshnyaga. Problem. Discrete Cosine Transform (DCT) is a very computationally expensive task. DCT require s a measurable amount of energy and many operations.

I am the owner, or an agent authorized to act on behalf of the owner, of the copyrighted work described.

## A Zero-Value Prediction Technique for Fast DCT Computation

Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author.While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server.

- - - - - - - - - - - - - - - - - - - - - - - - - - E N D - - - - - - - - - - - - - - - - - - - - - - - - - -
Presentation Transcript

### A Zero-Value Prediction Technique for Fast DCT Computation

Y. Nishida, K. Inoue, and V. Moshnyaga

Problem
• Discrete Cosine Transform (DCT) is a very

• DCT requires a measurable amount of energy

and many operations.

Reducing the number of operations and energy consumption of DCT becomes important.

Solution

A Zero-Value Prediction Technique for

Fast DCT Computation.

When the DCT outputs N zeros, we predict the remaining result is also zero.

Example of N=2

The detected string

of zeros

The predicted values

Starting point of

zero-value prediction

• Reduces the number of DCT and quantization

operations.

• Reduces energy consumption.
• DCT speed increases.

• The picture quality deteriorates a little.
Operation reduction rate as a function of N（DCT）

carphone

carphone

foreman

foreman

missa

missa

salesman

salesman

×

×

Operation reduction rate

Operation reduction rate

0.5

0.5

0.4

0.4

0.3

0.3

0.2

0.2

0.1

0.1

0

0

3

3

1

1

2

2

4

4

5

5

6

6

7

7

8

8

9

9

10

10

11

11

12

12

13

13

14

14

15

15

16

16

17

17

18

18

19

19

20

20

N

N

• The reduction rate is high when N is small.

carphone

carphone

foreman

foreman

missa

missa

salesman

salesman

×

×

PSNR as a function of N

70

60

50

PSNR (dB)

40

30

20

10

0

3

1

2

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

N

• The PSNR is low when N is small.
• When N=9, can reduce the number of DCT operations by 29%.
• The PSNR decreases by 1.47dB.