Search papers, labs, and topics across Lattice.
K Color Bits 5 4 3 2 1 5 4 3 2 1 5 4 3 2 1 Colors/Image 32 16 8 4 2 32 16 8 4 2 32 16 8 4 2 Prune Ratio 80% 83% 87.5% 92% 96% 80% 83% 87.5% 92% 96% 80% 83% 87.5% 92% 96% Random 88.15 84.38 80.15 77.04 70.08 57.36 50.96 42.19 39.71 36.68 62.54 58.18 53.34 50.32 20.28 Entropy 79.08 74.59 67.58 64.51 57.06 41.83 35.45 29.77 28.96 22.16 55.80 44.59 42.04 31.04 17.06 EL, N 89.81 88.47 87.01 74.77 73.17 61.83 60.75 52.16 32.07 28.17 64.71 58.15 56.54 45.19 29.14 TDDS (Strategy-E) 91.30 88.72 87.47 77.32 72.46 63.01 61.44 55.13 32.15 26.55 62.56 56.48 54.91 43.91 29.56 DCQ (Ours) 94.39 93.34 91.02 89.15 79.90 69.05 67.89 65.02 57.69 38.44 66.99 64.34 62.02 49.69 35.95 Full Accuracy 95.45% 78.21% 73.54% Table 2: Ablation Studies of Different Strategies., University of Electronic Science and Technology of China (UESTC)
1
0
3
Squeeze your image datasets without sacrificing model accuracy: DCQ intelligently reduces color redundancy, outperforming naive compression methods.
