Table 3. Statistics of glomerulus classification using different color texture descriptors (mean±std.).
Feature set | Dimension | FLD | SVM | KNN | |||
---|---|---|---|---|---|---|---|
ACC | AUC | ACC | AUC | ACC | AUC | ||
LBP [8] | 36 | 0.705±0.032 | 0.775±0.032 | 0.695±0.040 | 0.770±0.052 | 0.691±0.035 | 0.755±0.043 |
ind-LBP | 3×36 | 0.716±0.030 | 0.799±0.031 | 0.720±0.030 | 0.792±0.051 | 0.714±0.034 | 0.802±0.033 |
OC-LBP [25] | 6×36 | 0.730±0.028 | 0.813±0.025 | 0.730±0.038 | 0.830±0.052 | 0.716±0.031 | 0.803±0.032 |
norm-LBP [27] | 36 | 0.685±0.031 | 0.752±0.032 | 0.689±0.047 | 0.746±0.051 | 0.675±0.042 | 0.733±0.041 |
LCVBP [14] | 4×36 | 0.758±0.031 | 0.844±0.026 | 0.785±0.033 | 0.853±0.052 | 0.792±0.031 | 0.866±0.028 |
LCC [30] | 256 | 0.582±0.036 | 0.619±0.038 | 0.570±0.041 | 0.595±0.065 | 0.583±0.034 | 0.610±0.045 |
LCC-LBP [30] | 256+36 | 0.665±0.029 | 0.726±0.029 | 0.626±0.040 | 0.709±0.054 | 0.641±0.040 | 0.702±0.048 |
LAP [31] | 3×36 | 0.715±0.029 | 0.780±0.030 | 0.721±0.046 | 0.795±0.041 | 0.696±0.017 | 0.767±0.021 |
LAP-LBP [31] | 6×36 | 0.830±0.024 | 0.903±0.018 | 0.813±0.048 | 0.901±0.036 | 0.776±0.027 | 0.861±0.022 |
QDCP | 36 | 0.809±0.024 | 0.892±0.022 | 0.839±0.023 | 0.896±0.020 | 0.843±0.025 | 0.876±0.043 |
QDCP-LBP | 2×36 | 0.847±0.023 | 0.919±0.020 | 0.846±0.027 | 0.920±0.024 | 0.838±0.031 | 0.913±0.016 |
QDCP/LBP | 362 | 0.930±0.017 | 0.977±0.010 | 0.897±0.066 | 0.961±0.062 | 0.838±0.032 | 0.920±0.022 |