Image compressing apparatus

• US 5,845,016 A
• Filed: 06/27/1996
• Issued: 12/01/1998
• Est. Priority Date: 06/28/1995
• Status: Expired due to Term

• Current Assignee: Sharp Kabushiki Kaisha (Hon Hai Precision Industry Co., Ltd.)
• Original Assignee: Sharp Kabushiki Kaisha (Hon Hai Precision Industry Co., Ltd.)
• Inventors: Matsui, Hirofumi, Miyamoto, Masayuki, Iizuka, Kunihiko, Fujio, Mitsuhiko
• Primary Examiner(s): Mancuso, Joseph
• Assistant Examiner(s): AHMED, SAMIR ANWAR
• Application Number:
  US08/673,720
• Time in Patent Office: 887 Days
• Field of Search: 382/253, 382/232, 382/239, 348/401, 348/405, 348/414, 348/417, 348/418, 348/421, 348/422
• US Class Current: 382/253
• CPC Class Codes: G06T 9/008 Vector quantisation
• CPC Class Codes: H03M 7/3082 Vector coding for television signals, see H04N19/94
• CPC Class Codes: H04N 19/94 Vector quantisation

Abstract

An image compressing apparatus employs a mean-separated normalized vector quantization method according to which, with respect to vector components corresponding to input images inputted from image sensors via a plurality of lines, encodes and outputs a scalar-quantized code of a mean value, a scalar-quantized code of a maximum scalar product value with each code word in a code book, and an index of one of the code words which yields a maximum scalar product value. In this image compressing apparatus, when the maximum scalar product value is less than a predetermined threshold value, in accordance with judgement by a comparator circuit, an output selecting circuit stops outputting the codes of the maximum scalar product value and of the index, and outputs only the code of the mean value. Therefore, when the image is uniform with pixels varying little in their luminance levels in compression processing unit blocks, code data to be outputted are restricted so that only data of the mean value are outputted. Consequently, it is possible to restrain degradation of the image and to considerably reduce data amount.

Claims

1. An image compressing apparatus for receiving image signals as an input vector for every predetermined compression processing unit block and compressing the image signals with a mean-separated normalized vector quantization method, comprising:

a plurality of scalar product value calculating sections for calculating scalar product values of a difference component vector and a code word, the difference component vector being generated by subtracting a mean value of vector components of the input vector from each of the vector components, the code word being generated from image signals of a plurality of kinds of predetermined images for learning;

a scalar-quantization section for scalar-quantizing the mean value and a maximum scalar product value;

an encoding section for encoding an identification number of one of the code words, the code word yielding the maximum scalar product value through the calculation by said scalar product value calculating section;

a low band buffer for generating a low band vector for every low band compression processing unit block composed of a plurality of compression processing block units by receiving the mean values at a predetermined cycle;

a low band vector quantization section for vector-quantizing the low band vector with the mean-separated normalized vector quantization method;

a comparator section for judging whether the maximum scalar product value is not less than a predetermined threshold value; and

an output selecting section for, in accordance with the judgement by said comparator section, outputting the quantization code of the maximum scalar product value from said scalar quantization section, the identification code of the code word from said encoding section, a quantization code of a low band mean value, a quantization code of a low band maximum scalar product value and a low band identification code, i. e., an identification number of the code word yielding the low band maximum scalar product value when the maximum scalar product value is not less than the threshold value, and outputting the quantization code of the low band mean value, the quantization code of the low band maximum scalar product value, and the low band identification code,

wherein said output selecting section selectively outputs the quantization code of the mean value from said scalar quantization section, a quantization code of the low band mean value obtained by further vector-quantizing the mean value with the mean-separated normalized vector quantization method, the quantization code of the low band maximum scalar product value and the low band identification code.

2. The image compressing apparatus as defined in claim 1,

wherein said low band buffer includes:

a shift register for shifting a control signal and outputting the control signal as trigger signals in a sequential manner; and

a plurality of sample &

hold circuits for sampling and holding the mean value in accordance with the trigger signals from the shift register, and

said low band vector quantization section includes;

said scalar product value calculating sections;

said scalar quantization section;

said encoding section; and

said comparator section.

3. The image compressing apparatus as defined in claim 1,

wherein each of said scalar product value calculating sections includes an amplifier and capacitors corresponding to code components of each of the code words,

wherein said scalar product value calculating section calculates scalar product values of the input vector and each of the code words in a parallel manner for each of the code words by inputting input signals of a plurality of channels corresponding to the vector components to a terminal of each of the capacitors and connecting the other terminals of the capacitors commonly to the amplifier.

4. The image compressing apparatus as defined in claim 3,

wherein capacities of the capacitors are set in a capacity ratio as the quantized code components,

wherein each of said scalar product value calculation sections includes a correction capacitor corresponding to quantization error caused by the capacitor and inputs a product of a electrostatic capacity of the correction capacitor and the mean value to the amplifier as a correction value.

5. The image compressing apparatus as defined in claim 1, further comprising:

an input vector inversion selecting section for inverting polarity of the input vector and selectively changing-over obtained inverted and non-inverted values to be outputted; and

an inversion &

comparator section for, in accordance with the change-over operation of said input vector inversion selecting section, comparing the two scalar product values calculated by said scalar product value calculating section using the non-inverted and inverted values with each other, and outputting a greater scalar product value of the two and the code of the greater scalar product value.

6. The image compressing apparatus as defined in claim 1,

wherein the code word is one of a plurality of vectors for learning obtained in accordance with inputted image signals, the vector for learning having a power greater than a predetermined threshold value and great change in images in the compression processing unit blocks.

7. An image compressing apparatus for receiving image signals as an input vector for every predetermined compression processing unit block and compressing the image signals with a mean-separated normalized vector quantization method, comprising:

a mean value calculating section for calculating a mean value of vector components of the input vector;

a mean value separating section for generating a difference component vector by subtracting the calculated mean value from each of the vector components of the input vector;

a memory section for storing vectors for learning generated from image signals of a plurality of kinds of predetermined images for learning as code words;

a scalar product value calculating section for calculating scalar product values of the difference component vector and each of the code words;

a maximum scalar product value detecting section for detecting one of the code words, the code word yielding a maximum scalar product value through the calculation;

a mean value quantization section for scalar-quantizing the mean value;

a gain quantization section for scalar-quantizing the maximum scalar product value;

a vector encoding section for encoding an identification number of the code word yielding the maximum scalar product value;

a low band buffer for generating a low band vector for every low band compression processing unit block composed of a plurality of compression processing block units by receiving the mean values at a predetermined cycle;

a low band vector quantization section for vector-quantizing the low band vector with the mean-separated normalized vector quantization method;

a comparator section for judging whether the maximum scalar product value is not less than a predetermined threshold value; and

an output selecting section for, in accordance with the judgement by said comparator section, outputting the quantization code of the maximum scalar product value from said gain quantization section, the identification code of the code word from said vector encoding section, a quantization code of a low band mean value, a quantization code of a low band maximum scalar product value and a low band identification code, i. e., an identification number of the code word yielding the low band maximum scalar product value, instead of the quantization code of the mean value when the maximum scalar product value is not less than the threshold value, and outputting the quantization code of the low band mean value, the quantization code of the low band maximum scalar product value and the low band identification code when the maximum scalar product value is less than the threshold value,

wherein said output selecting section selectively outputs the quantization code of the mean value from said mean value quantization section, a quantization code of the low band mean value obtained by further vector-quantizing the mean value with the mean-separated normalized vector quantization method, the quantization code of the low band maximum scalar product value and the low band identification code.

8. The image compressing apparatus as defined in claim 7,

wherein said low band buffer includes:

a shift register for shifting a control signal and outputting the control signal as trigger signals in a sequential manner; and

a plurality of sample &

hold circuits for sampling and holding the mean value in accordance with the trigger signals from the shift register, and said low band vector quantization section includes;

said scalar product value calculating sections;

said scalar quantization section;

said encoding section; and

said comparator section.

9. The image compressing apparatus as defined in claim 7,

wherein each of said scalar product value calculating sections includes an amplifier and capacitors corresponding to code components of each of the code words,

wherein said scalar product value calculating section calculates scalar product values of the input vector and each of the code words in a parallel manner for each of the code words by inputting input signals of a plurality of channels corresponding to the vector components to a terminal of each of the capacitors and connecting the other terminals of the capacitors commonly to the amplifier.

10. The image compressing apparatus as defined in claim 9,

wherein capacities of the capacitors are set in a capacity ratio as the quantized code components,

wherein each of said scalar product value calculation sections includes a correction capacitor corresponding to quantization error caused by the capacitor and inputs a product of a electrostatic capacity of the correction capacitor and the mean value to the amplifier as a correction value.

11. The image compressing apparatus as defined in claim 7, further comprising:

an input vector inversion selecting section for inverting polarity of the input vector and selectively changing-over obtained inverted and non-inverted values to be outputted; and

an inversion &

comparator section for, in accordance with the change-over operation of said input vector inversion selecting section, comparing the two scalar product values calculated by said scalar product value calculating section using the non-inverted and inverted values with each other, and outputting a greater scalar product value of the two and the code of the greater scalar product value.

12. The image compressing apparatus as defined in claim 7,

wherein the code word is one of a plurality of vectors for learning obtained in accordance with inputted image signals, the vector for learning having a power greater than a predetermined threshold value and great change in images in the compression processing unit blocks.

13. An image compressing apparatus for receiving image signals as an input vector for every predetermined compression processing unit block and compressing the image signals with a mean-separated normalized vector quantization method, comprising:

a plurality of scalar product value calculating sections for calculating scalar product values of a difference component vector and a code word, the difference component vector being generated by subtracting a mean value of vector components of the input vector from each of the vector components, the code word being generated from image signals of a plurality of kinds of predetermined images for learning;

an input vector inversion selecting section for inverting polarity of the input vector and selectively changing-over obtained inverted and non-inverted values to be outputted;

an inversion &

comparator section for, in accordance with the change-over operation of said input vector inversion selecting section, comparing the two scalar product values calculated by said scalar product value calculating section using the non-inverted and inverted values with each other, and outputting a greater scalar product value of the two and the code of the greater scalar product value;

a scalar-quantization section for scalar-quantizing the mean value and a maximum scalar product value;

an encoding section for encoding an identification number of one of the code words, the code word yielding the maximum scalar product value through the calculation by said scalar product value calculating section;

a comparator section for judging whether the maximum scalar product value is not less than a predetermined threshold value; and

an output selecting section for, in accordance with the judgement by said comparator section, outputting the quantization code of the mean value from said scalar quantization section, the quantization code of the maximum scalar product value from said scalar quantization section and the identification code of the code word from said encoding section when the maximum scalar product value is not less than the threshold value, and outputting only the quantization code of the mean value when the maximum scalar product value is less than the threshold value.

14. An image compressing apparatus for receiving image signals as an input vector for every predetermined compression processing unit block and compressing the image signals with a mean-separated normalized vector quantization method, comprising:

a mean value calculating section for calculating a mean value of vector components of the input vector;

a mean value separating section for generating a difference component vector by subtracting the calculated mean value from each of the vector components of the input vector;

a memory section for storing vectors for learning generated from image signals of a plurality of kinds of predetermined images for learning as code words;

a scalar product value calculating section for calculating scalar product values of the difference component vector and each of the code words;

an input vector inversion selecting section for inverting polarity of the input vector and selectively changing-over obtained inverted and non-inverted values to be outputted;

an inversion &

comparator section for, in accordance with the change-over operation of said input vector inversion selecting section, comparing the two scalar product values calculated by said scalar product value calculating section using the non-inverted and inverted values with each other, and outputting a greater scalar product value of the two and the code of the greater scalar product value;

a maximum scalar product value detecting section for detecting one of the code words, the code word yielding a maximum scalar product value through the calculation;

a mean value quantization section for scalar-quantizing the mean value;

a gain quantization section for scalar-quantizing the maximum scalar product value;

a vector encoding section for encoding an identification number of the code word yielding the maximum scalar product value;

a comparator section for judging whether the maximum scalar product value is not less than a predetermined threshold value; and

an output selecting section for, in accordance with the judgement by said comparator section, outputting the quantization code of the mean value from said mean value quantization section, the quantization code of the maximum scalar product value from said gain quantization section and the identification code of the code word from said vector encoding section when the maximum scalar product value is not less than the threshold value, and outputting only the quantization code of the mean value when the maximum scalar product value is less than the threshold value.