|
09 May 2023, Volume 38 Issue 3
|
|
|
Abstract
In recent years, researchers have paid more and more attention to Indian handwritten text recognition. India is home to over 1600 distinct languages with 22 recognized as official. In the state of Karnataka, Kannada is the official language. When it comes to the number of native speakers in the world, Kannada comes in at number 33. Numerous scientists have worked to perfect the process of automating optical character recognition. However, the process of handwritten character recognition (HCR) has yet to be enhanced. This survey provides comprehensive guidance for automatic Kannada HCR and demonstrates the significant work needed to create a full HCR system. In this regard, the current survey study sheds light on the advancement of the appropriate techniques to fulfil the goal of building an automatic HCR. The study focuses on all of the steps needed in recognizing Kannada handwriting. The first step is to gather Kannada information that has been written by hand. The next step is to use various processing techniques such as binarization, normalization, noise rejection, morphological approaches, augmentation, picture enhancement, and skew correction to get rid of the stray data points. Character segmentation from Handwritten Kannada text is the third Step. Line, block-wise, and character-based segmentation are all part of the segmentation process. The next step is feature extraction, which takes the segmented character and extracts the salient features. The last step is to use Machine learning (ML) and deep learning (DL) for automatic character recognition from Kannada handwritten script. The purpose of this survey is to help emerging researchers by providing a narrative and detailed analysis of automatic Kannada Handwritten script recognition methods.
Keyword
Kannada, Character Recognition, Handwritten, Feature Extraction, Line Segmentation, Machine Learning, Accuracy.
PDF Download (click here)
|