IMAGE ENCODING METHOD, IMAGE DECODING METHOD, IMAGE ENCODING DEVICE AND IMAGE DECODING DEVICE

• US 20130176431A1
• Filed: 02/27/2013
• Published: 07/11/2013
• Est. Priority Date: 09/03/2010
• Status: Active Grant

• Current Assignee: Panasonic Semiconductor Solutions Co., Ltd. (Winbond Electronics Corporation)
• Original Assignee: Panasonic Corporation (Panasonic Holdings Corporation)
• Inventors: OGAWA, MAYU, KITAMURA, SHINJI
• Granted Patent: US 10,223,811 B2
• Time in Patent Office: Days
• US Class Current: 348/143
• CPC Class Codes: G06T 9/008 Vector quantisation
• CPC Class Codes: H04N 19/124 Quantisation
• CPC Class Codes: H04N 19/14 Coding unit complexity, e.g. amount of activity or edge presence estimation H04N19/146 takes precedence
• CPC Class Codes: H04N 19/182 the unit being a pixel
• CPC Class Codes: H04N 19/593 involving spatial prediction techniques
• CPC Class Codes: H04N 23/80 Camera processing pipelines; Components thereof

Abstract

In an image encoding device, a predicted pixel generator generates a predicted value for pixel data based on at least one of pixels adjacent to a pixel to be compressed. A quantization processor quantizes a difference value between the pixel data and the generated predicted value to a quantized value having a smaller number of bits than that of the pixel data to compress the pixel data to the quantized value. In a quantization width determiner, an edge determiner determines whether or not there are both a flat portion and an edge portion, based on a characteristic of the difference value in the group, and an edge pixel determiner determines whether each of pixels included in the group is a pixel in an edge portion or a pixel in a flat portion, based on the difference value, whereby a quantization width used in the quantization processor is determined.

Claims

1. An image encoding method for receiving pixel data of a pixel to be compressed, and compressing the input pixel data, comprising:

a predicted pixel generation step of generating a predicted value for the pixel data based on at least one of pixels adjacent to the pixel to be compressed;

a quantization step of quantizing a difference value between the pixel data and the predicted value generated in the predicted pixel generation step to a quantized value having a smaller number of bits than that of the pixel data, to compress the pixel data to the quantized value; and

a quantization width determination step of determining a quantization width used in the quantization step,

wherein

in the quantization width determination step, it is determined whether or not there are both a flat portion and an edge portion, in groups of a plurality of the pixels, and based on a result of the determination, quantization widths used in the quantization step are adaptively changed on a pixel-by-pixel basis.

2. The image encoding method of claim 1, wherein

the quantization width determination step includes

an edge determination step of determining whether or not there are both a flat portion and an edge portion, based on a characteristic of the difference value in the group, and

an edge pixel determination step of determining whether each of pixels included in the group is a pixel in an edge portion or a pixel in a flat portion, based on the difference value, and

based on a result of the determination of the edge determination step and a result of the determination of the edge pixel determination step, the quantization widths used in the quantization step are adaptively changed on a pixel-by-pixel basis.

3. The image encoding method of claim 2, wherein

in the edge determination step, if a difference between largest and smallest ones of the difference values in the group is larger than or equal to a first threshold, it is determined that there are both a flat portion and an edge portion.

4. The image encoding method of claim 3, wherein

the first threshold is a power of two.

5. The image encoding method of claim 2, wherein

in the edge determination step, if the largest one of the difference values in the group is larger than or equal to a second threshold and the smallest one of the difference values is smaller than or equal to a third threshold, it is determined that there are both a flat portion and an edge portion.

6. The image encoding method of claim 5, wherein

the second and third thresholds are each a power of two.

7. The image encoding method of claim 2, wherein

in the edge pixel determination step, it is determined that a pixel in the group is a pixel in an edge portion if the difference value is larger than or equal to a fourth threshold, and a pixel in a flat portion otherwise.

8. The image encoding method of claim 7, wherein

the fourth threshold is a power of two.

9. The image encoding method of claim 2, wherein

in the quantization width determination step, it is determined whether to use a first quantization width, or a second and a third quantization width, as the quantization width used in the quantization step, on a group-by-group basis, based on a result of the determination of the edge determination step, and it is determined whether to use the second quantization width or the third quantization width as the quantization width used in the quantization step, on a pixel-by-pixel basis, based on a result of the determination of the edge pixel determination step.

10. The image encoding method of claim 2, wherein

the quantization width determination step includes

an initial quantization width calculation step of calculating a first quantization width corresponding to the largest one of the difference values in the group, and

an edge portion quantization width calculation step of calculating a second quantization width used in quantization of an edge portion and a third quantization width used in quantization of a flat portion.

11. The image encoding method of claim 10, wherein

the second quantization width has a value obtained by adding one to the first quantization width.

12. The image encoding method of claim 10, wherein

the third quantization width is smaller than the first quantization width.

13. The image encoding method of claim 10, wherein

in the edge portion quantization width calculation step, a quantized width is selected from a plurality of the third quantization widths corresponding to a difference between the largest and smallest ones of the difference values in the group.

14. The image encoding method of claim 1, further comprising:

a packing step of inserting, into the quantized value, a 1-bit flag indicating whether or not a pixel is located in an edge portion, and outputting the resulting quantized value.

15. The image encoding method of claim 1, further comprising:

a packing step of generating, for each of the groups, S-bit encoded data containing at least one of the quantized value for pixels in the group, information about the quantization width, and a result of the determination of whether or not there are both a flat portion and an edge portion, where S is a natural number.

16. The image encoding method of claim 1, wherein

a difference value between input N-bit pixel data, where N is a natural number, and the predicted value having the same bit length, is encoded to a fixed-length code having a predetermined bit length of M, where M is a natural number smaller than N.

27. The image encoding method of claim 9, wherein

the quantization width determination step includes

an initial quantization width calculation step of calculating a first quantization width corresponding to the largest one of the difference values in the group, and

an edge portion quantization width calculation step of calculating a second quantization width used in quantization of an edge portion and a third quantization width used in quantization of a flat portion.

17. An image decoding method, wherein

input encoded data contains a quantized value, and

the quantized value is obtained by quantizing a difference value between pixel data and a predicted value for the pixel data, and the predicted value for the pixel data is generated based on at least one of pixels adjacent to a pixel of the pixel data, and

the quantized value is decoded to the pixel data, and

the method comprises:

a predicted pixel generation step of generating the predicted value for the current pixel data based on the past pixel data which has already been decoded;

an inverse quantization step of inverse quantizing the quantized value to generate the difference value having the quantized value which has been inverse quantized; and

a quantization width determination step of determining a quantization width used in the inverse quantization step, on a pixel-by-pixel basis, based on encoded quantization information and an encoded quantized value, and

in the quantization width determination step, quantization widths used in the inverse quantization step are adaptively changed on a pixel-by-pixel basis.

18. An image encoding device for receiving pixel data of a pixel to be compressed, and compressing the input pixel data, comprising:

a predicted pixel generator configured to generate a predicted value for the pixel data based on at least one of pixels adjacent to the pixel to be compressed;

a quantizer configured to quantize a difference value between the pixel data and the predicted value generated by the predicted pixel generator to a quantized value having a smaller number of bits than that of the pixel data to compress the pixel data to the quantized value; and

a quantization width determiner configured to determine a quantization width used in the quantizer,

wherein

the quantization width determiner determines whether or not there are both a flat portion and an edge portion, in groups of a plurality of the pixels, and based on a result of the determination, adaptively changes quantization widths used in the quantizer on a pixel-by-pixel basis.

20. The image encoding device of claim 18, wherein

the pixel data is RAW data input from an imaging element.

21. The image encoding device of claim 18, wherein

the pixel data is a luminance signal or a color difference signal which are generated by an image signal processing device based on RAW data which is input from an imaging element to the image signal processing device.

22. The image encoding device of claim 18, wherein

the pixel data is a luminance signal or a color difference signal which are obtained by a decompressing device decompressing a joint photographic experts group (JPEG) image.

23. A digital still camera comprising the image encoding device of claim 18, and comprising an image decoding device, wherein

input encoded data contains a quantized value, and

the quantized value is obtained by quantizing a difference value between pixel data and a predicted value for the pixel data, and the predicted value for the pixel data is generated based on at least one of pixels adjacent to a pixel of the pixel data, and

the quantized value is decoded to the pixel data, and

the image decoding device comprises:

a predicted pixel generator configured to generate the predicted value for the current pixel data based on the past pixel data which has already been decoded;

an inverse quantizer configured to inverse quantize the quantized value to generate the difference value having the quantized value which has been inverse quantized; and

a quantization width determiner configured to determine a quantization width used in the inverse quantizer, on a pixel-by-pixel basis, based on encoded quantization information and an encoded quantized value, and

the quantization width determiner adaptively changes quantization widths used in the inverse quantizer on a pixel-by-pixel basis.

24. A digital camcorder comprising the image encoding device of claim 18, and comprising an image decoding device, wherein

input encoded data contains a quantized value, and

the quantized value is obtained by quantizing a difference value between pixel data and a predicted value for the pixel data, and the predicted value for the pixel data is generated based on at least one of pixels adjacent to a pixel of the pixel data, and

the quantized value is decoded to the pixel data, and

the image decoding device comprises:

a predicted pixel generator configured to generate the predicted value for the current pixel data based on the past pixel data which has already been decoded;

an inverse quantizer configured to inverse quantize the quantized value to generate the difference value having the quantized value which has been inverse quantized; and

a quantization width determiner configured to determine a quantization width used in the inverse quantizer, on a pixel-by-pixel basis, based on encoded quantization information and an encoded quantized value, and

the quantization width determiner adaptively changes quantization widths used in the inverse quantizer on a pixel-by-pixel basis.

25. A surveillance camera comprising the image encoding device of claim 18, and comprising an image decoding device, wherein

input encoded data contains a quantized value, and

the quantized value is obtained by quantizing a difference value between pixel data and a predicted value for the pixel data, and the predicted value for the pixel data is generated based on at least one of pixels adjacent to a pixel of the pixel data, and

the quantized value is decoded to the pixel data, and

the image decoding device comprises:

a predicted pixel generator configured to generate the predicted value for the current pixel data based on the past pixel data which has already been decoded;

an inverse quantizer configured to inverse quantize the quantized value to generate the difference value having the quantized value which has been inverse quantized; and

a quantization width determiner configured to determine a quantization width used in the inverse quantizer, on a pixel-by-pixel basis, based on encoded quantization information and an encoded quantized value, and

the quantization width determiner adaptively changes quantization widths used in the inverse quantizer on a pixel-by-pixel basis.

26. A solid imaging element comprising:

the image encoding device of claim 18.

19. An image decoding device, wherein

input encoded data contains a quantized value, and

the quantized value is obtained by quantizing a difference value between pixel data and a predicted value for the pixel data, and the predicted value for the pixel data is generated based on at least one of pixels adjacent to a pixel of the pixel data, and

the quantized value is decoded to the pixel data, and

the device comprises:

a predicted pixel generator configured to generate the predicted value for the current pixel data based on the past pixel data which has already been decoded;

an inverse quantizer configured to inverse quantize the quantized value to generate the difference value having the quantized value which has been inverse quantized; and

a quantization width determiner configured to determine a quantization width used in the inverse quantizer, on a pixel-by-pixel basis, based on encoded quantization information and an encoded quantized value, and

the quantization width determiner adaptively changes quantization widths used in the inverse quantizer on a pixel-by-pixel basis.