An optimized pipelined architecture of LBT for JPEG XR encoding

  • Guanghui Ren*
  • , Xiaokai Hu
  • , Gangyi Wang
  • , Jinglong Wu
  • *Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

Abstract

JPEG XR is a new image coding standard that achieves high compression performance and low complexity. In this study, we propose a novel architecture of Lapped Biorthogonal Transform (LBT) for JPEG XR encoding using an optimized Quadriphase-RAM (QRAM) and a fully pipelined LBT module. In the previous architectures, the whole image data are stored in the main memory so that the waste of memory and the delay for the whole image acquisition must be inevitable. The proposed QRAM minimizes the memory size and improves the speed of image processing and the fully pipelined LBT module achieves higher throughput than that of the previous architectures. Results show the proposed architecture which achieves 121.2 M pixel/sec at 125 MHz can be used for the real-time JPEG XR encoding.

Original languageEnglish
Pages (from-to)1376-1382
Number of pages7
JournalInformation Technology Journal
Volume9
Issue number7
DOIs
StatePublished - 2010
Externally publishedYes

Keywords

  • FCT
  • FPGA
  • Full pipeline
  • Pre-filter
  • QRAM

Fingerprint

Dive into the research topics of 'An optimized pipelined architecture of LBT for JPEG XR encoding'. Together they form a unique fingerprint.

Cite this