1 / 39

LOW COMPLEXITY EMBEDDED QUANTIZATION SCHEME COMPATIBLE WITH BITPLANE IMAGE CODING

LOW COMPLEXITY EMBEDDED QUANTIZATION SCHEME COMPATIBLE WITH BITPLANE IMAGE CODING. Francesc Aulí -Llinàs. Department of Information and Communications Engineering Universitat Autònoma de Barcelona, Spain. TABLE OF CONTENTS. INTRODUCTION. 2-STEP SCALAR DEADZONE QUANTIZATION. α W. H. L.

Download Presentation

LOW COMPLEXITY EMBEDDED QUANTIZATION SCHEME COMPATIBLE WITH BITPLANE IMAGE CODING

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. LOW COMPLEXITY EMBEDDED QUANTIZATION SCHEMECOMPATIBLE WITH BITPLANE IMAGE CODING Francesc Aulí-Llinàs Department of Information and Communications Engineering UniversitatAutònoma de Barcelona, Spain

  2. TABLE OF CONTENTS INTRODUCTION 2-STEP SCALAR DEADZONE QUANTIZATION αW H L EXPERIMENTAL RESULTS CONCLUSIONS

  3. TABLE OF CONTENTS INTRODUCTION 2-STEP SCALAR DEADZONE QUANTIZATION αW H L EXPERIMENTAL RESULTS CONCLUSIONS

  4. INTRODUCTION QUALITY PROGRESSIVITY compressed codestream • Progressive transmission • Interactive applications • Codestream truncation • Image transcoding 1

  5. INTRODUCTION CLASSIC SCHEME: USDQ+BPC W = 24 0 10 1

  6. INTRODUCTION CLASSIC SCHEME: USDQ+BPC 10(10 emit 0 0 0 0 0 0 0 1 1 1 1 1 1 emit 1 W = 24 0 XXXX(2 23 1XXX(2 22 10XX(2 21 101X(2 20 1010(2 = 10(10 2

  7. INTRODUCTION CLASSIC SCHEME: USDQ+BPC 10(10 W = 24 0 XXXX(2 23 1XXX(2 22 10XX(2 21 101X(2 20 1010(2 = 10(10 2

  8. INTRODUCTION CLASSIC SCHEME: USDQ+BPC 10(10 W = 24 0 XXXX(2 23 1XXX(2 22 10XX(2 21 101X(2 20 1010(2 = 10(10 IS USDQ+BPC OPTIMAL FOR WAVELET-BASED LOSSY IMAGE CODING? density 2

  9. INTRODUCTION GENERAL EMBEDDED QUANTIZATION (GEQ) 10(10 emit 0 emit 0 0 0 0 0 0 1 1 emit 1 emit 1 1 1 1 W = 24 0 T1 1 yes > T1? no T2 T3 T4’ T4’’’ T4’’ yes > T4’’’? 0 no T5 T6’’’’’ T6’’’’’’ T6’’’’’’’ T6’’’ T6’ T6’’ T6’’’’ yes > T6’’’’’’ ? 0 no 3

  10. INTRODUCTION GENERAL EMBEDDED QUANTIZATION (GEQ) 10(10 W = 24 0 T1 1 yes > T1? no T2 T3 RESEARCH PURPOSE: ADAPT THE LOW-COMPLEXITY GEQ SCHEME TO BITPLANE CODING T4’ T4’’’ T4’’ yes > T4’’’? 0 no T5 T6’’’’’ T6’’’’’’ T6’’’’’’’ T6’’’ T6’ T6’’ T6’’’’ yes > T6’’’’’’ ? 0 no USDQ+BPC is optimal in terms of coding performance GEQ schemes can achieve same coding performance as that of USDQ+BPC employing fewer quantization stages GEQ schemes can help to reduce the computational costs of the codec in 20% GEQ is notcompatible with bitplane coding strategies 100 ≠ 10(10 3

  11. TABLE OF CONTENTS INTRODUCTION 2-STEP SCALAR DEADZONE QUANTIZATION αW H L EXPERIMENTAL RESULTS CONCLUSIONS 1

  12. 2-STEP SCALAR DEADZONE QUANTIZATION PROPOSED SCHEME W 0 USDQ+BPC same number of subintervals GEQ αW 2SDQ H L 1) (1 - α) L = CONDITIONS H α 2) 3) Each quantization stage halves the previous subintervals except in the first stage 4

  13. 2-STEP SCALAR DEADZONE QUANTIZATION PROPOSED SCHEME 10(10 emit 0 0 0 0 0 0 0 1 1 emit 1 1 1 1 1 W = 24 0 XXXX αW 1XXX 11XX 110X 1100 H L 2SDQ(1010(2)=1100 5

  14. 2-STEP SCALAR DEADZONE QUANTIZATION IMPLEMENTATION IN JPEG2000 ORIGINAL IMAGE 6

  15. 2-STEP SCALAR DEADZONE QUANTIZATION MULTI-COMPONENT TRANSFORM IMPLEMENTATION IN JPEG2000 ORIGINAL IMAGE 6

  16. 2-STEP SCALAR DEADZONE QUANTIZATION MULTI-COMPONENT TRANSFORM WAVELET TRANSFORM QUANTIZATION IMPLEMENTATION IN JPEG2000 ORIGINAL IMAGE 6

  17. 2-STEP SCALAR DEADZONE QUANTIZATION MULTI-COMPONENT TRANSFORM WAVELET TRANSFORM QUANTIZATION IMPLEMENTATION IN JPEG2000 ORIGINAL IMAGE 6

  18. 2-STEP SCALAR DEADZONE QUANTIZATION MULTI-COMPONENT TRANSFORM WAVELET TRANSFORM QUANTIZATION IMPLEMENTATION IN JPEG2000 ORIGINAL IMAGE 6

  19. 2-STEP SCALAR DEADZONE QUANTIZATION MULTI-COMPONENT TRANSFORM TIER-1 CODING TIER-2 CODING WAVELET TRANSFORM QUANTIZATION IMPLEMENTATION IN JPEG2000 ORIGINAL IMAGE JP2 CODESTREAM 6

  20. 2-STEP SCALAR DEADZONE QUANTIZATION MULTI-COMPONENT TRANSFORM TIER-1 CODING TIER-2 CODING WAVELET TRANSFORM QUANTIZATION IMPLEMENTATION IN JPEG2000 ORIGINAL IMAGE JP2 CODESTREAM 6

  21. 2-STEP SCALAR DEADZONE QUANTIZATION MULTI-COMPONENT TRANSFORM TIER-1 CODING TIER-2 CODING WAVELET TRANSFORM QUANTIZATION IMPLEMENTATION IN JPEG2000 ORIGINAL IMAGE JP2 CODESTREAM 6

  22. 2-STEP SCALAR DEADZONE QUANTIZATION MULTI-COMPONENT TRANSFORM TIER-1 CODING TIER-2 CODING WAVELET TRANSFORM QUANTIZATION IMPLEMENTATION IN JPEG2000 ORIGINAL IMAGE JP2 CODESTREAM 6

  23. 2-STEP SCALAR DEADZONE QUANTIZATION MULTI-COMPONENT TRANSFORM TIER-1 CODING TIER-2 CODING WAVELET TRANSFORM QUANTIZATION IMPLEMENTATION IN JPEG2000 ORIGINAL IMAGE JP2 CODESTREAM 7

  24. 2-STEP SCALAR DEADZONE QUANTIZATION MULTI-COMPONENT TRANSFORM TIER-1 CODING TIER-2 CODING WAVELET TRANSFORM QUANTIZATION IMPLEMENTATION IN JPEG2000 ORIGINAL IMAGE α2M 2M 0 2M-1 JP2 CODESTREAM 7

  25. 2-STEP SCALAR DEADZONE QUANTIZATION MULTI-COMPONENT TRANSFORM TIER-1 CODING TIER-2 CODING WAVELET TRANSFORM QUANTIZATION IMPLEMENTATION IN JPEG2000 ORIGINAL IMAGE α2M 2M 0 H L 2M-1 2SDQ header bit JP2 CODESTREAM 7

  26. 2-STEP SCALAR DEADZONE QUANTIZATION MULTI-COMPONENT TRANSFORM TIER-1 CODING TIER-2 CODING WAVELET TRANSFORM QUANTIZATION IMPLEMENTATION IN JPEG2000 ORIGINAL IMAGE α2M 2M 0 H L variable constant βH = 4(2αln2 – α + 1 – ln2) βL= 4α 2M-1 2SDQ header bit RD-opt JP2 CODESTREAM 2M-1 2M-2 0 7

  27. 2-STEP SCALAR DEADZONE QUANTIZATION MULTI-COMPONENT TRANSFORM TIER-1 CODING TIER-2 CODING WAVELET TRANSFORM QUANTIZATION IMPLEMENTATION IN JPEG2000 ORIGINAL IMAGE 2SDQ header bit RD-opt JP2 CODESTREAM 8

  28. 2-STEP SCALAR DEADZONE QUANTIZATION MULTI-COMPONENT TRANSFORM TIER-1 CODING TIER-2 CODING WAVELET TRANSFORM QUANTIZATION IMPLEMENTATION IN JPEG2000 ORIGINAL IMAGE 2SDQ header bit RD-opt JP2 CODESTREAM 8

  29. TABLE OF CONTENTS INTRODUCTION 2-STEP SCALAR DEADZONE QUANTIZATION αW H L EXPERIMENTAL RESULTS CONCLUSIONS 1

  30. EXPERIMENTAL RESULTS “Portrait” image (ISO 12640-1 corpus) Codeblocksof 64x64 2SDQ is applied on codeblockswith 6 bitplanes or more removing 1 bitplane 9

  31. EXPERIMENTAL RESULTS “Portrait” image (ISO 12640-1 corpus) Codeblocksof 64x64 2SDQ is applied on codeblockswith 6 bitplanes or more removing 1 bitplane 9

  32. EXPERIMENTAL RESULTS “Portrait” image (ISO 12640-1 corpus) Codeblocksof 64x64 2SDQ is applied on codeblockswith 6 bitplanes or more removing 1 bitplane 9

  33. EXPERIMENTAL RESULTS “Portrait” image (ISO 12640-1 corpus) Codeblocksof 64x64 2SDQ is applied on codeblockswith 6 bitplanes or more removing 1 bitplane 9

  34. EXPERIMENTAL RESULTS “Portrait” image (ISO 12640-1 corpus) Codeblocksof 64x64 2SDQ is applied on codeblockswith 6 bitplanes or more removing 1 bitplane 9

  35. EXPERIMENTAL RESULTS “Portrait” image (ISO 12640-1 corpus) Codeblocksof 64x64 2SDQ is applied on codeblockswith 6 bitplanes or more removing 1 bitplane 9

  36. EXPERIMENTAL RESULTS “Cafeteria” image (ISO 12640-1 corpus) Codeblocksof 64x64 2SDQ is applied on codeblockswith 6 bitplanes or more removing 1 bitplane 10

  37. TABLE OF CONTENTS INTRODUCTION 2-STEP SCALAR DEADZONE QUANTIZATION αW H L EXPERIMENTAL RESULTS CONCLUSIONS 1

  38. CONCLUSIONS Motivation explore new quantization schemes for wavelet-based image coding compatible with bitplane coding 2SDQ quantization scheme with 2 step sizes adapted to the density of wavelet coefficients αW Implementation replacement of USDQ quantization indices by 2SDQ indices H L Adaptation in JPEG2000 introduction of three easy-to-implement steps in the coding pipeline Results reduction of coding passes without penalizing coding performance 11

  39. CONCLUSIONS CONCLUSIONS Motivation explore new quantization schemes for wavelet-based image coding compatible with bitplane coding 2SDQ quantization scheme with 2 step sizes adapted to the density of wavelet coefficients αW Implementation replacement of USDQ quantization indices by 2SDQ indices H L Adaptation in JPEG2000 introduction of three easy-to-implement steps in the coding pipeline Results reduction of coding passes without penalizing coding performance 11

More Related