Image processing apparatus and method

• US 6,665,446 B1
• Filed: 12/20/1999
• Issued: 12/16/2003
• Est. Priority Date: 12/25/1998
• Status: Expired due to Term

• Current Assignee: Canon Kabushiki Kaisha (Canon Inc.)
• Original Assignee: Canon Kabushiki Kaisha (Canon Inc.)
• Inventors: Kato, Masami
• Primary Examiner(s): Boudreau, Leo
• Assistant Examiner(s): SHERALI, ISHRAT I
• Application Number:
  US09/466,907
• Time in Patent Office: 1,457 Days
• Field of Search: 382/232-233, 382/238-240, 382/243, 382/236, 382/251, 382/165, 382/250, 382/199, 382/224-225, 382/115, 382/118, 382/133, 382/282, 382/181, 375/305.199, 375/240.02, 375/240.12, 375/240.03, 375/240.11, 375/240.08, 375/240.18, 375/240.19, 375/240.2, 375/240.25, 358/518-520, 348/14.13, 348/155
• US Class Current: 382/251
• CPC Class Codes: G06T 9/001 Model-based coding, e.g. wire frame
• CPC Class Codes: H04N 19/124 Quantisation
• CPC Class Codes: H04N 19/17 the unit being an image region, e.g. an object
• CPC Class Codes: H04N 19/40 using video transcoding, i.e. partial or full decoding of a coded input stream followed by re-encoding of the decoded output stream
• CPC Class Codes: H04N 19/63 using sub-band based transform, e.g. wavelets

Abstract

A color-space conversion processor converts input digital image data into YUV color space, and a wavelet conversion processor performs wavelet conversion processing on the data. A face-area recognition processor extracts a face area in the image based on the obtained conversion coefficients. A quantization processor performs quantization processing on the input conversion coefficients while changing quantization coefficients used in the quantization processing in and out of the extracted face area. A variable-length coding processor encodes the quantized coefficients. A code synthesizing processor synthesizes the obtained respective color component code data with information indicating the face area extracted by the face-area recognition processor, and outputs the synthesized data as a code string.

Claims

1. An image processing apparatus comprising:

conversion means for converting input digital image data into coefficients based on spatial frequencies to obtain conversion coefficients;

extraction means for extracting a specific area in an image represented by lowest frequency components for a converted image obtained by said conversion means, said specific area being extracted from the converted image by pattern matching between an area having a predetermined color component value range and a predetermined shape; and

quantization means for performing quantization processing on said conversion coefficients of said converted image while changing a quantization characteristic in accordance with the specific area extracted by said extraction means.

2. The image processing apparatus according to claim 1, wherein said conversion means performs wavelet conversion on the input digital image data.

3. The image processing apparatus according to claim 2, wherein said extraction means extracts said specific area by using the lowest frequency component of the conversion coefficients obtained by said wavelet conversion.

4. The image processing apparatus according to claim 2, wherein said quantization means sets a quantization width for high frequency components as quantization coefficients for various frequency components obtained by said wavelet conversion, to be wider than that for the low frequency components, in said specific area and other areas than said specific area.

5. The image processing apparatus according to claim 2, wherein said quantization means uses a shortest quantization width for the lowest frequency component obtained by said wavelet conversion, whether a processed area is said specific area or not.

6. The image processing apparatus according to claim 1, wherein said extraction means extracts said specific area having pixels in which color difference component values have values in a predetermined range.

7. The image processing apparatus according to claim 6, wherein the color difference component values supplied to said extraction means are thinned in accordance with a predetermined reduction rate.

8. The image processing apparatus according to claim 1, wherein said extraction means performs the matching between a binary pattern representing positions of pixels having the color component values within the predetermined range in said input digital image data, and a binary pattern representing the predetermined shape, and extracts said specific area based on the result of said matching.

9. The image processing apparatus according to claim 8, wherein said predetermined shape is an elliptic shape, and wherein said specific area is an elliptic area.

10. The image processing apparatus according to claim 8, wherein said extraction means extracts said specific area from a position near the central portion of an image represented by said digital image data, prior to other portions.

11. The image processing apparatus according to claim 1, wherein said quantization means applies small quantization coefficients to the conversion coefficients within said specific area, while applies large quantization coefficients to the conversion coefficients out of said specific area.

12. The image processing apparatus according to claim 1, further comprising:

coding means for coding an output from said quantization means; and

output means for synthesizing code data obtained by said coding means with information indicating said specific area obtained by said extraction means, and outputting synthesized information.

13. The image processing apparatus according to claim 12, wherein the information indicating said specific area comprises a parameter representing a modeled area.

14. An image processing method comprising:

a conversion step of converting input digital image data into coefficients based on spatial frequencies to obtain conversion coefficients;

an extraction step of extracting a specific area in an image represented by lowest frequency components for a converted image obtained by the conversion step, said specific area being extracted from the converted image by pattern matching between an area having a predetermined color component value range and a predetermined shape; and

a quantization step of performing quantization processing on said conversion coefficients of the converted image while changing a quantization characteristic in accordance with the specific area extracted at said extraction step.

15. The image processing method according to claim 14, wherein at said conversion step, wavelet conversion is performed on the input digital image data.

16. The image processing method according to claim 15, wherein at said extraction step, said specific area is extracted by using the lowest frequency component obtained by said wavelet conversion.

17. The image processing method according to claim 15, wherein at said quantization step, a quantization width for high frequency components as quantization coefficients for various frequency components obtained by said wavelet conversion is set to be wider than that for the low frequency components, in said specific area and other areas than said specific area.

18. The image processing method according to claim 15, wherein at said quantization step, a smallest quantization width is used for the lowest frequency component obtained by said wavelet conversion, whether a processed area is said specific area or not.

19. The image processing method according to claim 14, wherein at said extraction step, said specific area, having pixels in which color difference component values have values in the predetermined range, is extracted.

20. The image processing method according to claim 19, wherein the color difference component values supplied to said extraction step are thinned in accordance with a predetermined reduction rate.

21. The image processing method according to claim 14, wherein at said extraction step, the matching is performed between a binary pattern representing positions of pixels having the color component values within the predetermined rate in said input digital image data, and a binary pattern representing the predetermined shape, and said specific area is extracted based on the result of said matching.

22. The image processing method according to claim 21, wherein said predetermined shape is an elliptic shape, and wherein said specific area is an elliptic area.

23. The image processing method according to claim 21, wherein at said extraction step, said specific area is extracted from a position near the central portion of an image represented by said digital image data, prior to other portions.

24. The image processing method according to claim 14, wherein at said quantization step, small quantization coefficients are applied to the conversion coefficients within said specific area, while large quantization coefficients are applied to the conversion coefficients out of said specific area.

25. The image processing method according to claim 14, further comprising:

a coding step of coding an output from said quantization means; and

an output step of synthesizing code data obtained at said coding step with information indicating said specific area obtained at said extraction step, and outputting synthesized information.

26. The image processing method according to claim 14, wherein the information indicating said specific area comprises a parameter representing a modeled area.

27. A storage medium containing a control program for causing a computer to compression-encode digital image data, said control program comprising:

code of conversion process for converting input digital image data into coefficients based on spatial frequencies to obtain conversion coefficients;

code of extraction process for extracting a specific area in an image represented by lowest frequency components for a converted image obtained by said conversion step, said specific area being extracted from the converted image by pattern matching between an area having a predetermined color component value range and a predetermined shape; and

code of quantization process for performing quantization processing on said conversion coefficients of the converted image while changing a quantization characteristic in accordance with the specific area extracted by said extraction means.