Size Invariant Handwritten Character Recognition using Single Layer Feedforward Backpropagation Neural Networks

Yousaf, Adeel; Khan, Muhammad Junaid; Khan, Muhammad Jaleed; Javed, Nizwa; Ibrahim, Haroon; Khurshid, Khurram

doi:10.1109/icomet.2019.8673459

Cited by 12 publications

(3 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The model was designed for easy testing and collaboration on a web server, utilizing the Google Colab cloud-based web interface for AI experiments. The paper [24] showed recognition of fixed-size handwritten characters using single-layer feed-forward neural networks. Their proposed system focused on offline recognition of numbers and the Latin alphabet.…”

Section: Literature Review and Problem Statementmentioning

confidence: 99%

Optimizing English character recognition: Enhanced accuracy through meta-heuristic algorithms

Al-Ghanimi

2024

Int. J. Circuit Comput. Networking

View full text Add to dashboard Cite

Recognition of symbols and words is important today, and neural networks can be used to detect these symbols and words. The main problem addressed is the requirement for a dependable and effective system that can achieve high accuracy despite the presence of varied font styles and minimal training data. Through the results that we present in this research, we conclude that the results obtained are better than the results reached in previous research methods by combining the Grasshopper Optimization algorithm and the Propeller algorithm. Specifically, compared with existing methods, our method improves the accuracy by about 1.11% to 6.35% for uppercase letters and 3.3% to 7.5% for lowercase letters. We can say that the Grasshopper algorithm helps in identifying the initial ideal areas Also, through an algorithm Propeller, we can improve these areas optimizes, Through this, we can determine the maximum possible degree of accuracy in the neural network. We can say that mixing these systems leads to a more powerful and effective system for character recognition, especially with differences in fonts. Through the results that we were able to obtain, we notice many features and characteristics that can help us solve the problem at hand: Firstly, very high accuracy in recognizing uppercase and lowercase letters in relation to the English language and comparing them with currently available methods. Secondly, a great ability to adapt to various font styles and also the ability to maintain high accuracy with the data specified for training. This proposed method can also recognize the handwriting of small and large letters in the English language, as these results showed great accuracy and efficiency in character recognition. Through this, it can be more reliable in real-world scenarios with different fonts and available data.

show abstract

Section: Literature Review and Problem Statementmentioning

confidence: 99%

Optimizing English character recognition: Enhanced accuracy through meta-heuristic algorithms

Al-Ghanimi

2024

Int. J. Circuit Comput. Networking

View full text Add to dashboard Cite

show abstract

“…Yousaf et al [16] present a methodology for handwritten English character (alphabetic and digital) preprocessing and recognition through an NN with one hidden layer, without using feature extraction. Preprocessing consists of binarization (by Otsu's global thresholding method), segmentation, filtering, and dynamic resizing.…”

Section: Related Workmentioning

confidence: 99%

“…The approach in [13] had an easier task to tackle: recognition of only lowercase letters. The approach in [16], although it declares 7+% better accuracy, it is a result of much easier recognition tasks: recognizing only capital letters or only digits. Finally, the approach in [17], although it shows about 5.5% better performance, it is also due to the much easier task of recognizing only digits and a much larger data set was used.…”

Section: Comparisonsmentioning

confidence: 99%

Optical Recognition of Handwritten Logic Formulas Using Neural Networks

et al. 2021

View full text Add to dashboard Cite

In this paper, we present a handwritten character recognition (HCR) system that aims to recognize first-order logic handwritten formulas and create editable text files of the recognized formulas. Dense feedforward neural networks (NNs) are utilized, and their performance is examined under various training conditions and methods. More specifically, after three training algorithms (backpropagation, resilient propagation and stochastic gradient descent) had been tested, we created and trained an NN with the stochastic gradient descent algorithm, optimized by the Adam update rule, which was proved to be the best, using a trainset of 16,750 handwritten image samples of 28 × 28 each and a testset of 7947 samples. The final accuracy achieved is 90.13%. The general methodology followed consists of two stages: the image processing and the NN design and training. Finally, an application has been created that implements the methodology and automatically recognizes handwritten logic formulas. An interesting feature of the application is that it allows for creating new, user-oriented training sets and parameter settings, and thus new NN models.

show abstract

An automated and efficient convolutional architecture for disguise-invariant face recognition using noise-based data augmentation and deep transfer learning

Khan

Siddiqui

et al. 2021

Vis Comput

View full text Add to dashboard Cite

Size Invariant Handwritten Character Recognition using Single Layer Feedforward Backpropagation Neural Networks

Cited by 12 publications

References 26 publications

Optimizing English character recognition: Enhanced accuracy through meta-heuristic algorithms

Optimizing English character recognition: Enhanced accuracy through meta-heuristic algorithms

Optical Recognition of Handwritten Logic Formulas Using Neural Networks

An automated and efficient convolutional architecture for disguise-invariant face recognition using noise-based data augmentation and deep transfer learning

Contact Info

Product

Resources

About