Ensemble deep transfer learning model for Arabic (Indian) handwritten digit recognition

Alkhawaldeh, Rami S.; Alawida, Moatsum; Alshdaifat, Nawaf Farhan Funkur; Alma’aitah, Wafa’ Za’al; Almasri, Ammar

doi:10.1007/s00521-021-06423-7

Cited by 23 publications

(4 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Conceptually the base model layer is part of feature learning and classification [84]. The main challenge in determining the base model layer that fits the background of the research case has to have different experimental approaches, as was done by [24], [25], [29], [35], [36] to produce the best model, when the transfer learning process is carried out from the pre-trained model to the base model layer.…”

Section: Pre-trained Modelmentioning

confidence: 99%

CNN Ensemble Learning Method for Transfer learning: A Review

Sulistya,

Br Bangun,

Tyas

2023

Ilk. J. Ilm.

View full text Add to dashboard Cite

This study provides a review of CNN's ensemble learning method for transfer learning by highlighting sections such as review studies, datasets, pre-trained models, transfer learning, ensemble learning, and performance. The results indicate that the trend of ensemble learning, transfer learning ensemble, and transfer learning is growing every year. In 2022, there will be 35 papers reviewed related to this topic in this study. Some datasets contain apparent information starting from the dataset name, total data points, dataset splitting, target dataset availability, and type classification. ResNet-50, VGG-16, InceptionV3, and VGG-19 are used in most papers as pre-trained models and transfer learning processes. 50 (90.1%) papers use ensemble learning, and 5 (9.1%) do without ensemble learning. The reviewed paper summarizes several performance measurements, including accuracy, precision, recall, f1score, sensitivity, specificity, training accuracy, validation accuracy, test accuracy, training losses, validation losses, test losses, training time, and AUC, DSC. In the last section, 49 papers produce the best model performance using the proposed model, and 6 other papers use DenseNet, DeQueezeNet, Extended Yager Model, InceptionV3,

show abstract

Section: Pre-trained Modelmentioning

confidence: 99%

CNN Ensemble Learning Method for Transfer learning: A Review

Sulistya,

Br Bangun,

Tyas

2023

Ilk. J. Ilm.

View full text Add to dashboard Cite

show abstract

“…Finally, this study was only limited to six variables, but in future work it is recommended to include other factors that could affect the users' behavioral intentions to use AMAN MH-App in Jordan, such as technology Anxiety, trust, compatibility, Self-Efficacy, Risk, and perceived trust. Also, this study suggested to use Machine Learning techniques (Almasri, Alkhawaldeh, and Celebi 2020;Almasri, Celebi, and Alkhawaldeh 2019) and Deep Learning methods (Alkhawaldeh 2022) to predict the user's intention to use mobile health Apps.…”

Section: Limitation and Future Workmentioning

confidence: 99%

The Users’ Behavioral Intention to use Mobile Health-Tech Application to Prevent the Spreading of Coronavirus

Almasri

2022

South East European Journal of Economics and Business

Self Cite

View full text Add to dashboard Cite

The emergence of mobile health applications (MH-Apps) has enhanced the healthcare field’s services, particularly in the treatment, diagnosis, and follow-up. AMAN Mobile Health Application (AMAN MH-App) is one of the health-tech solutions used to fight the Coronavirus pandemic. It has a built-in feature to track users’ activities to protect users from contacting an infected person. However, the acceptance of AMAN MH-App in Jordan is still in an early stage, and the number of users has reached 15% of the country’s population. Therefore, this study aims to assess the use of AMAN MH-App among young people using the quantitative method. A total of (450) valid samples participated in the study after removing 33 invalid samples. Smart-PLS 3.2.7 was used for data analysis. The findings showed that all independent variables (Perceived usefulness, Perceived ease of use value, Subjective norms, Perceived behavioral control, Information Credibility, and Optimism) positively impact on the dependent variable (Behavioural Intention to use AMAN MH-App). We believe that AMAN MH-App’s information’s credibility (i.e. providing up-to-date, authoritative, accurate, and trustworthy information) will increase the number of the App’s users. The results of this research can be applied to similar context and applications in different countries.

show abstract

“…There are a total of ten possible digits, in addition to special symbols such as delimiters and number separators. Although recognizing isolated digits may be a relatively easier problem with published works reporting accuracies above 99% (e.g., [16,17]), recognizing the complete courtesy amount is a difficult problem because of a number of reasons apart from the common reasons associated with any handwriting recognition tasks, such as variability in pen-strokes and writing style. In the case of courtesy amount, writers optionally add some special delimiter symbols at the beginning and end of the amounts.…”

Section: Introductionmentioning

confidence: 99%

A Hybrid Rule-Based and Machine Learning System for Arabic Check Courtesy Amount Recognition

Ahmad

2023

Sensors

View full text Add to dashboard Cite

Courtesy amount recognition from bank checks is an important application of pattern recognition. Although much progress has been made on isolated digit recognition for Indian digits, there is no work reported in the literature on courtesy amount recognition for Arabic checks using Indian digits. Arabic check courtesy amount recognition comes with its own unique challenges that are not seen in isolated digit recognition tasks and, accordingly, need specific approaches to deal with them. This paper presents an end-to-end system for courtesy amount recognition starting from check images as input to recognizing amounts as a sequence of digits. The system is a hybrid system, combining rule-based modules as well as machine learning modules. For the amount recognition system, both segmentation-based and segmentation-free approaches were investigated and compared. We evaluated our system on the CENPARMI dataset of real bank checks in Arabic. We achieve 67.4% accuracy at the amount level and 87.15% accuracy at the digit level on the test set consisting of 626 check images. The results are presented with detailed analysis, and some possible future work is identified. This work can be used as a baseline to benchmark future research in Arabic check courtesy amount recognition.

show abstract

Ensemble deep transfer learning model for Arabic (Indian) handwritten digit recognition

Cited by 23 publications

References 34 publications

CNN Ensemble Learning Method for Transfer learning: A Review

CNN Ensemble Learning Method for Transfer learning: A Review

The Users’ Behavioral Intention to use Mobile Health-Tech Application to Prevent the Spreading of Coronavirus

A Hybrid Rule-Based and Machine Learning System for Arabic Check Courtesy Amount Recognition

Contact Info

Product

Resources

About