Table 10 Comparison of proposed method with notable studies
From: Segmentation-free optical character recognition for printed Urdu text
Study | Dataset | Recognition unit | Accuracy | Complete ligature recognition |
|---|---|---|---|---|
Ahmed et al. [22] | UPTI | Character | 89.00% | – |
Hassan et al. [16] | UPTI | Character | 87.40%/ | – |
| Â | Â | Â | 94.85% | Â |
Naz et al. [17] | UPTI | Character | 96.40% | – |
Naz et al. [18] | UPTI | Character | 94.97% | – |
Javed et al. [10] | CLE | Ligature | 92.00% | No |
Akram et al. [13] | CLE | Ligature | 97.87% | No |
Javed and Hussain [19] | CLE | Ligature | 92.73% | No |
Khattak et al. [29] | CLE | Ligature | 97.93% | No |
Sabbour and Shafait [5] | UPTI | Ligature | 91.00% | No |
Akram et al. [28] | CLE | Ligature | 86.15% | Yes |
Hussain et al. [21] | CLE | Ligature | 87.76% | Yes |
Proposed | UPTI | Ligature | 92.26% | Yes |