稀疏分解和图拉普拉斯正则化的图像前景背景分割方法
|
谭婷芳,蔡万源,蒋俊正
|
Image foreground-background segmentation method based on sparse decomposition and graph Laplacian regularization
|
Tingfang TAN,Wanyuan CAI,Junzheng JIANG
|
|
表 1 不同方法在合成图像上的分割性能对比 |
Tab.1 Comparison of segmentation performance of different methods on synthetic image % |
|
方法 | 图1的第1张测试图 | | 图1的第2张测试图 | | | 平均值 | | P | R | F1 | | P | R | F1 | | P | R | F1 | LRSD | 73.11 | 71.49 | 72.29 | | 84.57 | 83.98 | 84.27 | | 78.84 | 77.74 | 78.28 | LAD | 80.83 | 82.08 | 81.45 | | 83.13 | 79.64 | 81.35 | | 81.98 | 80.86 | 81.40 | SDTVM | 85.99 | 82.63 | 84.28 | | 88.33 | 75.04 | 81.14 | | 87.16 | 82.63 | 82.71 | SR | 72.26 | 77.92 | 74.99 | | 85.39 | 88.14 | 86.74 | | 78.83 | 83.03 | 80.87 | SD-GFT | 98.33 | 82.97 | 90.00 | | 92.12 | 86.69 | 89.33 | | 95.83 | 84.83 | 83.24 | SDGLR | 99.45 | 84.34 | 91.28 | | 99.78 | 85.52 | 92.10 | | 99.62 | 84.93 | 91.69 |
|
|
|