An integer IDCT opencl implementation method and system conforming to JPEG standard

By generating integer DCT coefficient blocks on the CPU and copying them in batches to the GPU for parallel integer IDCT calculation, the problem of low data transfer and floating-point operation efficiency in JPEG decoding on general-purpose GPUs is solved, achieving efficient and compatible JPEG decoding.

CN122420548APending Publication Date: 2026-07-17COMP APPL TECH INST OF CHINA NORTH IND GRP
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
COMP APPL TECH INST OF CHINA NORTH IND GRP
Filing Date
2026-04-23
Publication Date
2026-07-17

Smart Images

  • Figure CN122420548A_ABST
    Figure CN122420548A_ABST
Patent Text Reader

Abstract

本发明涉及一种符合JPEG标准的整数IDCT OpenCL实现方法及系统，属于图像编码解码技术领域，解决了现有方法在通用GPU上存在数据搬运开销过大、浮点运算效率低下及缺乏针对现有JPEG解码平台的协同调度机制的技术问题。包括在CPU端对待解码JPEG图像的码流数据进行熵解码与反量化，生成多个整数型DCT系数块依次存入位于主机内存中的环形缓冲区中；将环形缓冲区中的多个整数型DCT系数块，通过PCIe总线批量拷贝至GPU显存中的全局输入缓冲区；在GPU端基于OpenCL配置，对每个整数型DCT系数块执行整数IDCT计算，得到符合JPEG标准的8‑bit像素值的解码块；将所有解码块存入GPU显存中的全局输出缓冲区，形成完整的解码图像。降低了PCIe数据搬运开销和浮点运算依赖，提升了通用GPU上的JPEG解码效率。
Need to check novelty before this filing date? Find Prior Art