Image encoding apparatus and method for the same and image decoding apparatus and method for the same

An image encoding apparatus, comprising:

• a pixel bit depth increase unit to convert bit depth of each pixel of an input image formed of a plurality of pixels each having an N bit depth into an (N+M) bit depth larger than the N bit depth by M bits;
• a prediction image generator to generate a prediction image of the (N+M) bit depth corresponding to the input image of the (N+M) bit depth from a reference image of the (N+M) bit depth;
• a subtracter to calculate a differential signal of between the input image of the (N+M) bit depth and the prediction image of the (N+M) bit depth;
• an encoding unit to output encoded image information by encoding the differential signal;
• a decoding unit to output a decoded difference image based on the encoded image information;
• an adder to add the prediction image of the (N+M) bit depth to the decoded difference image and output a decoded image of the (N+M) bit depth;
• a pixel bit depth converter to convert the bit depth of each pixel of the decoded image of the (N+M) bit to the N bit depth selectively using one of a plurality of mutually different conversion system;
• a reference image storing memory to store the decoded image converted into the N bit depth as the reference image;
• and
• a pixel bit depth inverse converter to convert the bit depth of each pixel of the reference image of the N bit depth read out from the reference image storing memory into the (N+M) bit depth larger than the N bit depth by M bits in accordance with the conversion system, wherein the pixel bit depth converter includes:
• a plurality of pixel bit depth converter units to convert a first local signal in units of one block of the decoded image of the (N+M) bit depth into a plurality of second local signals of the N bit depth with a plurality of mutually different conversion systems,
• a calculator to calculate a conversion system based on the first local signal, and
• a selector to output the second local signal selected based on the calculated conversion system and an additional information containing the signal of the conversion system for the second signal as the decoded image of the N bit depth.

An image decoding apparatus, comprising:

• a decoding unit to output a decoded difference image on receiving an encoded image information of an (N+M) bit depth;
• a prediction image generator to generate a prediction image of the (N+M) bit depth by a reference image of the (N+M) bit depth using the encoded image information;
• an adder to add the prediction image of the (N+M) bit depth to the decoded difference image and output a decoded image of the (N+M) bit depth;
• a pixel bit depth converter to convert each pixel of the decoded image of the (N+M) bit into the N bit depth selectively using one of a plurality of mutually different conversion systems;
• a reference image storing memory to store the decoded image converted into the N bit depth as the reference image;
• and
• a pixel bit depth inverse converter to convert the bit depth of each pixel of the reference image of the N bit depth into the (N+M) bit depth larger than N bit depth by M bits in accordance with the conversion system, wherein the pixel bit depth converter includes:
• a plurality of pixel bit depth converter units to convert a first local signal in units of one block of the decoded image of the (N+M) bit depth into a plurality of second local signals of the N bit depth with a plurality of mutually different conversion systems,
• a plurality of pixel bit depth inverse converter units to convert the second local signals into a plurality of third local signals of the (N+M) bit depth, respectively,
• subtracters to calculate a differential signal between the first local signal and the third local signal, and
• a selector to output the second local signal selected based on the differential signal calculated by the subtractor and an additional information containing the conversion system of the second local signal as the decoded image of the N bit depth.

An image encoding method, comprising:

• converting a bit depth of each pixel of an input image formed of a plurality of pixels each having an N bit depth into an (N+M) bit depth larger than the N bit depth by M bits;
• generating a prediction image of the (N+M) bit depth corresponding to the input image of the (N+M) bit depth by a reference image of the (N+M) bit depth;
• obtaining a differential signal between the input image of the (N+M) bit depth and the prediction image of the (N+M) bit depth;
• outputting an encoded image information by encoding the differential signal;
• outputting a decoded difference image based on the encoded image information;
• adding the prediction image of the (N+M) bit depth to the decoded difference image and outputting a decoded image of the (N+M) bit depth;
• converting the bit depth of each pixel of the decoded image of the (N+M) bit into the N bit depth selectively using one of a plurality of mutually different conversion systems;
• storing the decoded image converted into the N bit depth as the reference image;
• and
• converting the bit depth of each pixel of the reference image of the N bit depth read out from the reference image storing memory into the (N+M) bit depth larger than N bit depth by M bits in accordance with the conversion system, wherein
• the converting the bit depth of each pixel of the decoded image of the (N+M) bit into the N bit depth includes:
• converting a first local signal in units of one block of the decoded image of the (N+M) bit depth into a plurality of second local signals of the N bit depth with a plurality of mutually different conversion system,
• converting the second local signals into a plurality of third local signals of the (N+M) bit depth, respectively,
• calculating a differential signal between the first local signal and the third local signal, and
• outputting the second local signal selected based on the differential signal and an additional information containing the conversion system of the second local signal as the decoded image of the N bit depth.

An image decoding method, comprising:

• outputting a decoded difference image on receiving an encoded image information of an (N+M) bit depth;
• generating a prediction image of the (N+M) bit depth by a reference image of the (N+M) bit depth using the encoded image information;
• adding the prediction image of the (N+M) bit depth to the decoded difference image and outputting a decoded image of the (N+M) bit depth;
• converting the bit depth of each pixel of the decoded image of the (N+M) bit into the N bit depth selectively using one of a plurality of mutually different conversion systems;
• storing the decoded image converted into the N bit depth as the reference image;
• and
• converting the bit depth of each pixel of the reference image of the N bit depth into the (N+M) bit depth larger than N bit depth by M bits in accordance with the conversion system, wherein
• the converting the bit depth of each pixel of the decoded image of the (N+M) bit into the N bit depth includes:
• converting a first local signal in units of one block of the decoded image of the (N+M) bit depth into a plurality of second local signals of the N bit depth with a plurality of mutually different conversion systems,
• converting the second local signals into a plurality of third local signals of the (N+M) bit depth, respectively,
• calculating a differential signal between the first local signal and the third local signal, and
• outputting the second local signal selected based on the differential signal and an additional information containing the conversion system of the second local signal as the decoded image of the N bit depth.