Author Affiliations
1Microelectronics Research and Development Center, Shanghai University, Shanghai 200072, China2School of Mechatronic Engineering and Automation, Shanghai University, Shanghai 200072, Chinashow less
Fig. 1. 19-bit-plane image features of Lena
Fig. 2. 19-bit-plane CR distribution of Lena
Fig. 3. Distribution of field of view test points
Fig. 4. Image for exploring luminance masking effect
Fig. 5. Image for exploring contrast masking effect
Fig. 6. Asymmetric foveated concave fitting ellipse
Fig. 7. Subjective experimental results for luminance masking experiment. (a) Relationship between JND and eccentricity;
Fig. 8. Subjective experimental results for contrast masking experiment. (a) Relationship between JND and contrast; (b) relationship among JND, contrast and luminance; (c) JND value distribution in foveated range
Fig. 9. Noise injection distribution of three kinds of JND models. (a) Original image; (b) model in Ref. [
12]; (c) model in Ref. [
11];
Fig. 10. Enlarged view of central area after noise pollution for three JND models. (a) Image after noise pollution; (b) model in Ref. [
12]; (c) model in Ref. [
11]; (d) our model
Fig. 11. Hardware system. (a) Overall framework; (b) picture of hardware test platform
Fig. 12. Comparison of 19-bit-plane compression ratio distribution of Lena before and after algorithm processing
Fig. 13. Compression effect comparison of images and partially enlarged view. (a) Original image; (b) model in Ref. [
11]; (c) our model
Parameter | Value |
---|
Resolution(RGB)/pixel | 2560×2560 | Pixel per degree /pixel | 22 | Field angle /(°) | 120 |
|
Table 1. Performance parameters of VR glasses based on OLED-on-silicon micro-display
Attribute | Luminance masking | Contrast masking |
---|
Background luminance | 0,10,20,60,100,140,180,200,220,230 | 100,140,180,220 | Contrast luminance | 0 | 0,20,40,60,80 | Eccentricity /(°) | 1,2,3,5,10,15,20,25,30,35,40,45,50,55,60 | 1,2,3,5,10,15,20,25,30,35 |
|
Table 2. Stimulus attributes for luminance masking and contrast masking experiments
Image | PSNR /dB | MOS |
---|
Model in Ref.[12] | Model in Ref.[11] | Our model |
---|
Rapids | 58.44 | 4.2 | 4.8 | 4.9 | Bikes | 59.00 | 4.3 | 4.9 | 4.9 | Buildings | 57.66 | 4.2 | 4.7 | 4.9 | Carnivaldolls | 53.46 | 4.0 | 4.6 | 4.8 | Caps | 59.66 | 4.3 | 4.9 | 5.0 | Churchandcapitol | 55.57 | 4.2 | 4.8 | 4.9 | Coinsinfountain | 57.03 | 4.2 | 4.8 | 4.9 | Dancers | 55.11 | 4.4 | 5.0 | 5.0 | House | 59.12 | 4.3 | 4.9 | 4.9 | Lighthouse | 56.12 | 4.2 | 4.8 | 5.0 | Average | 57.107 | 4.24 | 4.82 | 4.91 |
|
Table 3. Comparison of MOSs of JND models under same PSNR
Bit-plane | BIT1 | BIT2 | BIT3 | BIT4 |
---|
Origin CR | 124.09 | 123.99 | 123.46 | 114.91 | Processed CR | 25.00 | 19.26 | 55.37 | 74.23 |
|
Table 4. Comparison of low 4-bit-plane compression ratio of Lena before and after algorithm processing
Image | PSNR(eX=0°)/dB | PSNR(eX=60°)/dB | FPSNR /dB | SSIM | CR/% |
---|
Ref.[11] | Ours | Ref. [11] | Ours | Ref. [11] | Ours | Ref.[11] | Ours | Ref. [11] | Ours |
---|
Bikes | 76.06 | 56.35 | 47.84 | 53.72 | 43.56 | 43.57 | 0.95 | 0.99 | 70.16 | 46.18 | Buildings | 84.44 | 62.48 | 47.38 | 58.55 | 42.99 | 45.01 | 0.96 | 0.99 | 66.52 | 51.60 | Caps | — | 55.77 | 47.26 | 51.65 | 43.18 | 43.75 | 0.91 | 0.98 | 68.84 | 31.61 | Carnivaldolls | 70.01 | 65.42 | 47.05 | 54.39 | 40.50 | 47.34 | 0.87 | 0.99 | 62.44 | 37.08 | Coinsinfountain | 86.72 | 56.50 | 48.08 | 55.05 | 42.24 | 43.65 | 0.95 | 0.99 | 64.42 | 41.64 | House | — | 64.17 | 47.93 | 56.16 | 43.08 | 47.66 | 0.93 | 0.99 | 67.80 | 36.31 | Lighthouse | 77.00 | 63.30 | 48.67 | 56.63 | 44.41 | 43.78 | 0.94 | 0.99 | 70.78 | 38.90 | Monarch | 66.94 | 59.59 | 50.26 | 56.40 | 42.61 | 45.13 | 0.90 | 0.98 | 67.67 | 35.68 | Ocean | — | 56.45 | 46.73 | 54.91 | 42.49 | 43.81 | 0.91 | 0.98 | 68.86 | 35.44 | Parrots | — | 56.06 | 47.78 | 54.52 | 42.88 | 43.34 | 0.89 | 0.98 | 67.29 | 38.08 | Plane | 65.47 | 64.51 | 47.89 | 50.96 | 42.76 | 46.49 | 0.92 | 0.99 | 60.63 | 29.68 | Rapids | 86.47 | 56.57 | 46.84 | 53.99 | 43.01 | 43.92 | 0.91 | 0.99 | 69.58 | 36.05 | Sailing4 | 74.96 | 56.56 | 48.37 | 56.28 | 43.33 | 43.57 | 0.92 | 0.99 | 69.14 | 37.15 | Stream | — | 56.45 | 47.93 | 54.70 | 43.48 | 43.89 | 0.97 | 0.99 | 68.23 | 48.47 | Studentsculpture | 86.20 | 62.55 | 47.77 | 56.52 | 42.13 | 47.04 | 0.97 | 0.99 | 63.80 | 49.73 | Average | 77.427 | 59.515 | 47.852 | 54.962 | 42.483 | 44.797 | 0.927 | 0.99 | 67.077 | 39.573 |
|
Table 5. Comparison of evaluation indexes and compression ratio of different compression algorithms for color image compression
Algorithm | PSNR /dB | SSIM | CR /% |
---|
Ref.[20] | 37.66 | 0.97 | 72.22 | Ours | 55.63 | 0.99 | 39.57 |
|
Table 6. Evaluation index and compression ratio comparison of bit-plane compression algorithms