Fig 1.
Overview of the proposed method.
Fig 2.
Contour curve fragments of the arabic letter ain.
Fig 3.
How to compute CPCA feature attributes at point Pi.
Fig 4.
How to compute CON3 feature attributes at point Pi.
Fig 5.
Writer identification with respect to contour fragment length on IAM and KHATT datasets.
Fig 6.
Writer identification with respect to GAP parameter on IAM and KHATT datasets.
Fig 7.
Writer identification with respect to angular quantization on IAM and KHATT datasets.
Fig 8.
Writer identification with respect to line segment length on IAM and KHATT datasets.
Fig 9.
Writer identification with respect to distance quantization parameters Nq and Nl on IAM and KHATT datasets.
Fig 10.
Writer identification with respect to Ø angle quantization on IAM and KHATT datasets.
Table 1.
The impacts of the available handwritten text.
Fig 11.
Writer identification with respect to the number of writes using NN classifier on IAM and KHATT datasets.
Fig 12.
Writer identification with respect to the number of writes using SVM classifier on IAM and KHATT datasets.
Fig 13.
Writer identification with respect to the codebook size using NN classifier on IAM and KHATT datasets.
Fig 14.
Writer identification with respect to the codebook size using SVM classifier On IAM and KHATT datasets.
Table 2.
CPCA results on IAM dataset.
Table 3.
CON3 results on IAM dataset.
Table 4.
COMBINED FEATURES results on IAM dataset.
Table 5.
CPCA results on KHATT dataset.
Table 6.
CON3 results on KHATT dataset.
Table 7.
COMBINED FEATURES results on KHATT dataset.
Table 8.
Comparison to the state-of-the-art writer identification methods for IAM dataset.
Table 9.
Comparison to the state-of-the-art writer identification methods for KHATT dataset.
Fig 15.
Proposed system misclassification errors.
Fig 16.
(a). The runtime of different system steps on IAM dataset using CON3 Feature. (b). The runtime of different system steps on IAM dataset using CPCA Feature.
Fig 17.
(a). The runtime of different system steps on KHATT dataset using CON3 Feature. (b). The runtime of different system steps on KHATT dataset using CPCA Feature.