| 摘要 |
<p>The present invention relates to video image processing. An image processing method for RDO based adaptive spatial-temporal resolution frame is provided, comprising: A. computing and obtaining a condition value of a critical bit-rate by an encoder based on the input video image and a change of the target allocated bit-rate; B. comparing the condition value of the critical bit-rate with a threshold value by the encoder, and obtaining an image resolution appropriate for the target allocated bit-rate via down-sampling if the condition value of critical bit-rate is lower than the threshold, i.e., the target allocated bit-rate is lower than the critical bit-rate, otherwise performing a general encoding process; C. recovering an original resolution of the received image by a decoder via up-sampling and smoothing it. The method of the present invention may obviously reduce distortion of the decoded video image, which is resulted from over-compression or over-quantization by the encoder in low bit-rate.</p> |
