西安电子科技大学学报 ›› 2016, Vol. 43 ›› Issue (4): 75-80.doi: 10.3969/j.issn.1001-2400.2016.04.014

• 研究论文 • 上一篇    下一篇

JPEG-LS近无损图像编码器VLSI结构设计

聂永康;雷杰;李云松;宋长贺;吴宪云   

  1. (西安电子科技大学 综合业务网理论及关键技术国家重点实验室,陕西 西安  710071)
  • 收稿日期:2015-04-27 出版日期:2016-08-20 发布日期:2016-10-12
  • 通讯作者: 雷杰(1981-),男,副教授,E-mail: jielei@mail.xidian.edu.cn
  • 作者简介:聂永康(1990-),男,西安电子科技大学硕士研究生,E-mail: 251782250@qq.com.
  • 基金资助:

    国家优秀青年基金资助项目(61222101);国家自然科学基金资助项目(61301287,61301291);111基地资助项目(B08038);中央高校基本科研业务费专项资金资助项目(K5051301043);中国博士后科学基金资助项目(2013M540735)

VLSI design of the JPEG-LS near-lossless image encoder

NIE Yongkang;LEI Jie;LI Yunsong;SONG Changhe;WU Xianyun   

  1. (State Key Lab. of Integrated Service Networks, Xidian Univ., Xi'an  710071, China)
  • Received:2015-04-27 Online:2016-08-20 Published:2016-10-12

摘要:

标准JPEG-LS图像压缩算法性能优越、算法复杂度低,因而被广泛应用于空间通信、医疗等领域.但是,该算法工作在近无损压缩模式时,计算当前像素的预测值需要使用前一像素的重建值,而重建值需要复杂运算才能得到,导致编码器存在大延时的反馈电路,这给编码器的高速硬件实现带来困难.在对JPEG-LS算法的深入研究的基础上,采用重建值的前向预测方法,提出一种全新的JPEG-LS近无损图像编码器超大规模集成电路结构,能够有效避免编码过程中的反馈延时.与现有方案相比,提出的超大规模集成电路结构具有编码速率高、不需要片外缓存、系统功耗低等优点,适用于资源受限的星载图像压缩.

关键词: 图像压缩, 近无损, JPEG-LS, 前向预测, 现场可编程门阵列

Abstract:

The JPEG-LS image compression standard brings excellent performance and low complexity, which makes it the widely used in space communications, medicine and many other fields. However, the prediction of the current pixel requires the reconstruction value of the previous pixel in the near-lossless mode, which leads to complicated computation and a feedback circuit with long latency in hardware implementations. On the basis of an intensive study of the JPEG-LS algorithm, a forward prediction method of reconstruction value is used and the new VLSI implementation of the JPEG-LS encoder is proposed in this paper, which can avoid the feedback latency. Compared with existing implementations, the proposed one has the advantages of higher encoding speed, no nuse of off-chip storages, and lower power consumption, making it suitable for resource-limited spaceborne image compression.

Key words: image compression, near-lossless, JPEG-LS, forward prediction, FPGA

Baidu
map