Table 4 Details of performance metrics used in the proposed study

Recognition accuracy (A)

\(\frac{\mathrm{TP}+\mathrm{TN}}{\mathrm{TP}+\mathrm{FP}+\mathrm{TN}+\mathrm{FN}}\)

It counts the correct predictions out of total predictions

Precision (P)

\(\frac{\mathrm{TP}}{\mathrm{TP}+\mathrm{FP}}\)

It counts the correct positive predictions out of the total positive predictions for given numeral class

Recall (R)

\(\frac{\mathrm{TP}}{\mathrm{TP}+\mathrm{FN}}\)

It counts the correct positive predictions out of the total samples for given numeral class

F1 score (F1)

\(2\times \frac{{\mathrm{P}}_{ }\times {\mathrm{R}}_{ }}{{\mathrm{P}}_{ }+{\mathrm{R}}}\)

It counts harmonic mean of precision and recall for given class