An Ensemble Learning Approach using Decision Fusion for the Recognition of Arabic Handwritten Characters

Dhief, Rihab; Youssef, Rabaa; Bouchemha, Amel

doi:10.5220/0010839500003122

Cited by 1 publication

(2 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The model is based on handcrafted features such as projections, profiles, widths and heights, extrema, concave arcs, endpoints, holes, and junctions [39]. The extracted features were then classified using KNN, SVM, and RF.…”

Section: Handwritten Arabic Recognitionmentioning

confidence: 99%

“…However, it achieved the second-best accuracy on the AHCD dataset. [32] 2017 94.90 [45] 2017 94.8 [34] 2021 92.5 [45] 2017 97.60 [44] 2014 94.28 [36] 2022 91.2 [46] 2021 98.21 [47] 2021 93.30 [37] 2022 91.0 [34] 2021 95.4 [48] 2021 95.20 This paper 2024 93.05 [43] 2021 97.00 [39] 2022 92.91 [49] 2022 96.78 [40] 2023 95.00 [36] 2022 98.48 This paper 2024 96.88 [37] 2022 98.00 [38] 2024 95.89 This paper 2024 98.30 6…”

Section: Comparison With Existing Approachesmentioning

confidence: 99%

See 1 more Smart Citation

Optimised CNN Architectures for Handwritten Arabic Character Recognition

Alghyaline

2024

CMC

View full text Add to dashboard Cite

Handwritten character recognition is considered challenging compared with machine-printed characters due to the different human writing styles. Arabic is morphologically rich, and its characters have a high similarity. The Arabic language includes 28 characters. Each character has up to four shapes according to its location in the word (at the beginning, middle, end, and isolated). This paper proposed 12 CNN architectures for recognizing handwritten Arabic characters. The proposed architectures were derived from the popular CNN architectures, such as VGG, ResNet, and Inception, to make them applicable to recognizing character-size images. The experimental results on three well-known datasets showed that the proposed architectures significantly enhanced the recognition rate compared to the baseline models. The experiments showed that data augmentation improved the models' accuracies on all tested datasets. The proposed model outperformed most of the existing approaches. The best achieved results were 93.05%, 98.30%, and 96.88% on the HIJJA, AHCD, and AIA9K datasets.

show abstract

Section: Handwritten Arabic Recognitionmentioning

confidence: 99%