Table 1

Confusion matrix for the classification results
PlanktoVision
det ukw 1 2 3 4 5 6 7 8 9 10 % recognition % false positive
det 11061 0 0 1 0 0 0 0 0 0 0 0 99.99 0.00
ukw 0 371 0 2 0 4 3 0 7 3 0 13 92.06 8.09
1 0 3 91 0 0 0 0 0 1 0 1 0 94.79 6.59
2 0 5 3 182 0 0 0 0 2 0 0 1 94.30 3.85
3 0 1 0 0 57 1 1 0 0 0 0 0 95.00 0.00
4 0 0 0 2 0 115 0 0 0 0 0 0 98.29 8.70
5 0 1 0 0 0 0 124 2 8 3 0 0 89.86 4.03
6 0 7 0 2 0 0 0 94 0 0 1 0 90.38 2.13
7 0 2 1 0 0 0 0 0 253 2 1 2 96.93 7.51
8 0 0 0 0 0 0 0 0 0 59 3 0 95.16 13.56
9 0 1 0 0 0 5 0 0 0 0 127 0 95.49 4.72
10 0 10 2 0 0 0 1 0 1 0 0 240 94.49 6.25
Ø 94.73 5.45

The data were classified into detritus (det), unknown plankton organism (ukw), Cyclotella (1), Anabeana (2), Chlorogonium (3), Cryptomonas (4), Desmodesmus (5), Staurastrum (6), Botryococcus (7), Pediastrum (8), Trachelomonas (9) and Crucigenia (10). The test set included images of 4 different samples that were independently prepared, imaged and analyzed on different days to prevent the selection of an over fitted classifier. The results of PlanktoVision (columns) were compared to a manual classification (rows). Correctly classified results are shown in bold.

Schulze et al.

Schulze et al. BMC Bioinformatics 2013 14:115   doi:10.1186/1471-2105-14-115

Open Data