Arabic Handwritten Letters Dataset (AHLD)

Citation Author(s):
Boutkhil
SIDAOUI
Marwa Djihane
BOUSSAHRA
Ikram
AMARA
Submitted by:
boutkhil sidaoui
Last updated:
Sat, 07/06/2024 - 07:08
DOI:
10.21227/qg80-7d57
License:
0
0 ratings - Please login to submit your rating.

Abstract 

Arabic handwritten letters Dataset (AHLD) consists of 8,000 handwritten Arabic letter images of size 128x128 pixels, distributed into 28 classes (Arabic alphabets). This dataset is derived from processing 582 images, each containing several letters, 

written by 15 individuals. The dataset creation involves a series of image processing operations: image acquisition, grayscale conversion,  binarization, noise reduction, segmentation, normalization, skeletonization, and dataset labeling. 

The final AHLD dataset was divided into two parts: the first represents 70% of the total dataset and will be used for training, and the second represents the remaining 30% of the dataset and will be used for testing.

Total dataset: 8000 images.

Training set: 70% of the dataset (5601 images).

Testing set: 30% of the dataset (2399 images).

Number of classes (labels): 28 Arabic alphabet letters.

Image size: 128x128 pixels. 

Instructions: 

This file contains description of the AHLD dataset

Comments

This dataset is created to help researchers in arabic handwritten recognition

Submitted by boutkhil sidaoui on Sat, 07/06/2024 - 07:12