发明名称 IMAGE ENCODING APPARATUS, IMAGE ENCODING METHOD, IMAGE DECODING APPARATUS, AND IMAGE DECODING METHOD
摘要 The present invention provides an image encoding/decoding technique that is capable of achieving the higher compression efficiency. An image encoding method comprises: an intra prediction step which performs intra prediction on a block basis to generate a predicted image; a subtraction step which calculates the difference in prediction between the predicted image generated by the intra prediction step and an original image; a frequency conversion step which performs frequency conversion processing for the difference in prediction; a quantization step which subjects the output of the frequency conversion step to quantization processing; and a variable-length encoding step which subjects the output of the quantization step to variable-length encoding processing; wherein the intra prediction encoding step predicts a target pixel to be encoded by use of pixel values of two reference pixels between which the target pixel to be encoded is located.
申请公布号 US2016366407(A1) 申请公布日期 2016.12.15
申请号 US201615245435 申请日期 2016.08.24
申请人 HITACHI MAXELL, LTD. 发明人 Takahashi Masashi;Murakami Tomokazu
分类号 H04N19/11;H04N19/159;H04N19/567;H04N19/124;H04N19/182;H04N19/17;H04N19/176 主分类号 H04N19/11
代理机构 代理人
主权项 1. An image decoding method that decodes an image by performing prediction processing, comprising: a prediction mode selection step, which selects one of a plurality of prediction modes for the prediction processing; and a prediction step, which performs the prediction processing of the prediction mode selected at the prediction mode selection step, wherein the prediction modes selectable at the prediction mode selection step include: a first intra prediction mode in which first reference pixels are generated employing pixels within a decoded block adjacent to one of left and upper sides of a target block to be decoded, and prediction pixels are generated through computation employing values of decoded second reference pixels adjacent to one of the upper and left sides of the target block to be decoded which is a different side from the first reference pixels and values of the first reference pixels; and a second intra prediction mode which is different from the first intra prediction mode and in which prediction pixels are generated employing values of the decoded pixels adjacent to the target block to be decoded, wherein there is a state in the first intra prediction mode in which calculation processing of values of the prediction pixels employing the computation is feasible for any pixels excepting all boundary pixels located at a lower end of the target block to be decoded within the target block to be decoded, wherein the prediction pixels are generated through computation employing interpolation processing between the values of the second reference pixels and the values of the first reference pixels in the first intra prediction mode.
地址 OSAKA JP