A Survey of Handwritten Character Recognition with MNIST and EMNIST
Open accessThe DOI matches its Crossref record; 84 of the 84 references it cites were found in a registry or an index.
Abstract
This paper summarizes the top state-of-the-art contributions reported on the MNIST dataset for handwritten digit recognition. This dataset has been extensively used to validate novel techniques in computer vision, and in recent years, many authors have explored the performance of convolutional neural networks (CNNs) and other deep learning techniques over this dataset. To the best of our knowledge, this paper is the first exhaustive and updated review of this dataset; there are some online rankings, but they are outdated, and most published papers survey only closely related works, omitting most of the literature. This paper makes a distinction between those works using some kind of data augmentation and works using the original dataset out-of-the-box. Also, works using CNNs are reported separately; as they are becoming the state-of-the-art approach for solving this problem. Nowadays, a significant amount of works have attained a test error rate smaller than 1% on this dataset; which is becoming non-challenging. By mid-2017, a new dataset was introduced: EMNIST, which involves both digits and letters, with a larger amount of data acquired from a database different than MNIST’s. In this paper, EMNIST is explained and some results are surveyed.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Figures
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Although the images are not of a real-world environment, they can be an appropriate alternative considering that the main objective of this study was to explore the possibility of deep learning for the objective quantification of the visual perception of the substituted signal.”
From Deep Learning-Based Optimization of Visual–Auditory Sensory Substitution (2023) · cites this paper as [31]
Experimental — accuracy not yet measured
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Gradient-based learning applied to document recognition
Lecun, Y., Bottou, L., Bengio, Y., & Haffner, P. (1998) · Proceedings of the IEEE · doi:10.1109/5.726791 · Cited by 58,805 · cited 7×
“For instance, Belongie et al. [5] achieved 0.63% and Keysers et al. [6] achieved 0.54% and 0.52% using K-NN, Kégl and Busa-Fekete [7] achieved 0.87% using boosted stumps on Haar features, LeCun et al. [4] achieved 0.8% and Decoste and Schölkopf [8] attained results from 0.56 to 0.68% using SVM.”
BackgroundMethodInfluentialExtreme learning machine: Theory and applications
Huang, G.-B., Zhu, Q.-Y., & Siew, C.-K. (2006) · Neurocomputing · doi:10.1016/j.neucom.2005.12.126 · Cited by 13,285 · cited 1×
“In 2015, McDonnell et al. [21] proposed an approach using the ‘extreme learning machine’ (ELM) [22] algorithm to train shallow non-convolutional neural networks, and they agree on the fact that data augmentation with distortions can reduce error rates even further.”
MethodShape matching and object recognition using shape contexts
Belongie, S., Malik, J., & Puzicha, J. (2002) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/34.993558 · Cited by 6,306 · cited 2×
“For instance, Belongie et al. [5] achieved 0.63% and Keysers et al. [6] achieved 0.54% and 0.52% using K-NN, Kégl and Busa-Fekete [7] achieved 0.87% using boosted stumps on Haar features, LeCun et al. [4] achieved 0.8% and Decoste and Schölkopf [8] attained results from 0.56 to 0.68% using SVM.”
MethodInfluentialDynamic Routing Between Capsules
Sabour, S., Frosst, N., & Hinton, G. E. (2017) · arXiv · doi:10.48550/arxiv.1710.09829 · Cited by 4,167 · cited 1×
“An interesting work using capsule layers (a concept recently introduced by Sabour et al. [76]) is TextCaps [77], a CNN with 3 convolutional layers and two capsule layers, which obtain accuracies of 95.36% in EMNIST Letters and 99.79% in EMNIST Digits.”
MethodInfluentialMulti-column deep neural networks for image classification
Ciresan, D., Meier, U., & Schmidhuber, J. (2012) · 2012 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2012.6248110 · Cited by 3,720
Best practices for convolutional neural networks applied to visual document analysis
Simard, P. Y., Steinkraus, D., & Platt, J. C. (2005) · Seventh International Conference on Document Analysis and Recognition, 2003. Proceedings. · doi:10.1109/icdar.2003.1227801 · Cited by 2,741 · cited 2×
“87% NN 2-layer 800 hidden units, MSE loss [9] 0.”
BackgroundSalakhutdinov, R., & Hinton, G. E. (2009) · International Conference on Artificial Intelligence and Statistics · Cited by 2,605
Lee, H., Grosse, R., Ranganath, R., & Ng, A. Y. (2009) · Proceedings of the 26th Annual International Conference on Machine Learning · doi:10.1145/1553374.1553453 · Cited by 2,412 · cited 2×
“Lee et al. [54] described convolutional deep belief networks with probabilistic max-pooling achieving an error rate of 0.82%.”
BackgroundInfluentialWhat is the best multi-stage architecture for object recognition?
Jarrett, K., Kavukcuoglu, K., Ranzato, M. A., & LeCun, Y. (2009) · 2009 IEEE 12th International Conference on Computer Vision · doi:10.1109/iccv.2009.5459469 · Cited by 2,153 · cited 3×
“…[39] 0.47% CNN (3 conv, 1 dense), stochastic pooling [25] 0.49% CNN (2 conv, 1 dense, relu) with dropout [24] 0.52% CNN, unsup pretraining [17] 0.53% CNN (2 conv, 1 dense, relu) with DropConnect [24] 0.57% SparseNet + SVM [15] 0.59% CNN (2 conv, 1 dense), unsup pretraining [16] 0.60%…”
MethodInfluential(2013) · Cited by 2,056
Research works printed but not linked: 5
- 23. Kasun, L.L.C.; Zhou, H.; Huang, G.B.; Vong, C.M. Representational learning with extreme learning machine for big data. IEEE Intell. Syst. 2013 , 28 , 31-34.
- 39. Lin, M.; Chen, Q.; Yan, S. Network In Network. In Proceedings of the 2nd International Conference on Learning Representations, Banff, AB, Canada, 14-16 April 2014.
- 42. Calderón, A.; Roa-Valle, S.; Victorino, J. Handwritten digit recognition using convolutional neural networks and Gabor filters. In Proceedings of the 2003 International Conference on Computational Intelligence, Cancun, Mexico, 19-21 May 2003.
- 73. Peng, Y.; Yin, H. Markov Random Field Based Convolutional Neuralx Networks for Image Classification. In IDEAL 2017: Intelligent Data Engineering and Automated Learning ; Lecture Notes in Computer Science; Yin, H., Gao, Y., Chen, S., Wen, Y., Cai, G., Gu, T., Du, J., Tallón-Ballesteros, A., Zhang, M., Eds.; Springer: Guilin, China, 2017; Volume 10585, pp. 387-396.
- 75. Mor, S.S.; Solanki, S.; Gupta, S.; Dhingra, S.; Jain, M.; Saxena, R. Handwritten text recognition: With deep learning and Android. Int. J. Eng. Adv. Technol. 2019 , 8 , 172-178.
Cited by
Arabic handwriting recognition system using convolutional neural network
Altwaijry, N., & Al-Turaiki, I. (2021) · Neural Computing and Applications · doi:10.1007/s00521-020-05070-8 · Cited by 228
Sakai, A., Komatsu, M., Komatsu, R., Matsuoka, R., Yasutomi, S., Dozen, A., Shozu, K., Arakaki, T., Machino, H., Asada, K., et al. (2022) · Biomedicines · doi:10.3390/biomedicines10030551 · Cited by 68
“The splitting ratio of the data was standard for machine learning [47–49].”
MethodA Hybrid Sparrow Search Algorithm of the Hyperparameter Optimization in Deep Learning
Fan, Y., Zhang, Y., Guo, B., Luo, X., Peng, Q., & Jin, Z. (2022) · Mathematics · doi:10.3390/math10163019 · Cited by 59
Classifying Garments from Fashion-MNIST Dataset Through CNNs
Henrique, A. S., Fernandes, A. M. D. R., Lyra, R., Leithardt, V. R. Q., Correia, S. D., Crocker, P. A., & Dazzi, R. L. S. (2021) · Advances in Science Technology and Engineering Systems Journal · doi:10.25046/aj0601109 · Cited by 50
Startups and Consumer Purchase Behavior: Application of Support Vector Machine Algorithm
Ebrahimi, P., Salamzadeh, A., Soleimani, M., Khansari, S. M., Zarea, H., & Fekete‐Farkas, M. (2022) · Big Data and Cognitive Computing · doi:10.3390/bdcc6020034 · Cited by 49
“Unsupervised learning is the path to realizing “real” artificial intelligence [60].”
BackgroundTwo Decades of Bengali Handwritten Digit Recognition: A Survey
Rahman, A., Hasan, M. B., Ahmed, S., Ahmed, T., Ashmafee, M. H., Kabir, M. R., & Kabir, M. H. (2022) · IEEE Access · doi:10.1109/access.2022.3202893 · Cited by 54
“The purpose of HDR systems is to encode handwritten digits (0-9 in English) into a computer interpretable format for creating a digital footprint of the information [5], [6].”
BackgroundInfluentialShao, J., Li, Z., Sun, W., Zhou, T., Sun, Y., Liu, L., Lin, Z., Mao, Y., & Zhang, J. (2026) · ACM Computing Surveys · doi:10.1145/3844940 · Cited by 55
“Our analysis primarily focuses on the most common privacy attacks in the literature, including model inversion attacks and membership inference attacks.”
Unsupervised feature learning-based encoder and adversarial networks
Suryawati, E., Pardede, H. F., Zilvan, V., Ramdan, A., Krisnandi, D., Heryana, A., Yuwana, R. S., Kusumo, R. B. S., Arisal, A., & Supianto, A. A. (2021) · Journal Of Big Data · doi:10.1186/s40537-021-00508-9 · Cited by 44
Alheraki, M., Al-Matham, R. N., & Al-Khalifa, H. S. (2023) · Human-Centric Intelligent Systems · doi:10.1007/s44230-023-00024-4 · Cited by 43
“Automatic handwritten recognition has been extensively investigated using various machine learning methods, including K-nearest Neighbors (KNNs) [3], Support Vector Machines (SVMs), transfer learning [4, 5], and deep learning techniques such as Neural Networks (NNs) [6].”
MethodZakariyya, I., Kalutarage, H., & Al-Kadri, M. O. (2023) · Computers & Security · doi:10.1016/j.cose.2023.103388 · Cited by 46
A Robust Model for Handwritten Digit Recognition using Machine and Deep Learning Technique
Agrawal, A. K., Shrivas, A. K., & Awasthi, V. K. (2021) · International Conference for Emerging Technology (INCET) · doi:10.1109/incet51464.2021.9456118 · Cited by 41
Handwritten Character Recognition on Android for Basic Education Using Convolutional Neural Network
Zin, T. T., Thant, S., Pwint, M. Z., & Ogino, T. (2021) · Electronics · doi:10.3390/electronics10080904 · Cited by 39
“The authors described an exhaustive review and updated survey with state-of-the-art methods on two well-known datasets, MNIST and EMNIST in [7].”
BackgroundDeep Convolutional Spiking Neural Networks for Image Classification
Vaila, R. (2021) · ScholarWorks (Boise State University) · doi:10.18122/td.1782.boisestate · Cited by 41
Threats on Machine Learning Technique by Data Poisoning Attack: A Survey
Ahmed, I. M., & Kashmoola, M. Y. (2021) · Communications in computer and information science · doi:10.1007/978-981-16-8059-5_36 · Cited by 50
Cho, Y., Hong, S., Kim, M., & Kim, J. (2022) · Electronics · doi:10.3390/electronics11162604 · Cited by 38
Yang, J., Shokouhifar, M., Por, L. Y., Khan, A. A., Awais, M., & Mousavi, Z. (2023) · Expert Systems with Applications · doi:10.1016/j.eswa.2023.122704 · Cited by 40
Vaila, R., Chiasson, J., & Saxena, V. (2020) · IEEE Transactions on Emerging Topics in Computational Intelligence · doi:10.1109/tetci.2020.3035164 · Cited by 44
Orr, W., & Crawford, K. (2024) · New Media & Society · doi:10.1177/14614448241251797 · Cited by 50
“They may be perceived as “solved” as models achieve high prediction scores (e.g. Baldominos et al., 2019).”
Manuscripts Character Recognition Using Machine Learning and Deep Learning
Islam, M. A., & Iacob, I. E. (2023) · Modelling—International Open Access Journal of Modelling in Engineering Science · doi:10.3390/modelling4020010 · Cited by 39
“A number of other studies [12–14, 16, 30– 35] used the data augmentation technique to overcome the data deficiency problem and to get commendable accuracy in the task of character recognition or classification.”
ResultA bibliometric analysis of off-line handwritten document analysis literature (1990–2020)
Ruiz-Parrado, V., Heradio, R., Aranda‐Escolástico, E., Sánchez, Á., & Vélez, J. F. (2021) · Pattern Recognition · doi:10.1016/j.patcog.2021.108513 · Cited by 31
Kim, E. K., Kim, J. Y., Lee, H., & Kim, S. (2021) · Applied Sciences · doi:10.3390/app11125586 · Cited by 34
“Data augmentation improves the generalization capabilities of the deep learning network and the performance of the classification model [17,18].”
BackgroundBradley, H., Louis, S., Trevillian, C., Quach, L., Bankowski, E. N., Slavin, A. N., & Tiberkevich, V. S. (2023) · AIP Advances · doi:10.1063/5.0128530 · Cited by 30
Convolutional Extreme Learning Machines: A Systematic Review
Rodrigues, I. R., Neto, S. R. D. S., Kelner, J., Sadok, D. F. H., & Endo, P. T. (2021) · Informatics · doi:10.3390/informatics8020033 · Cited by 28
“Similar to the object recognition problem, the handwritten digit and character recognition problem recurs in digital image processing and pattern recognition benchmarking [73].”
BackgroundResultShin, J., Maniruzzaman, M., Uchida, Y., Hasan, M. A. M., Megumi, A., Suzuki, A., & Yasumura, A. (2022) · Applied Sciences · doi:10.3390/app12105256 · Cited by 26
Francis, S. A., & Sangeetha, M. (2025) · Results in Control and Optimization · doi:10.1016/j.rico.2025.100532 · Cited by 32
Byun, U., Na, H., & Kim, S. (2025) · Advanced Functional Materials · doi:10.1002/adfm.202519431 · Cited by 27
COVID-19’s influence on cardiac function: a machine learning perspective on ECG analysis
Gomes, J. C., de Santana, M. A., Masood, A. I., de Lima, C. L., & dos Santos, W. P. (2023) · Medical & Biological Engineering & Computing · doi:10.1007/s11517-023-02773-7 · Cited by 26
“We adopted the 5-layer version of LeNet, trained considering the MNIST dataset [8, 36–38].”
MethodHuruf: An Application for Arabic Handwritten Character Recognition Using Deep Learning
Kamal, M., Shaiara, F., Abdullah, C. M., Ahmed, S., Ahmed, T., & Kabir, M. H. (2022) · International Conference on Computer and Information Technology (ICCIT) · doi:10.1109/iccit57492.2022.10054769 · Cited by 23
Cho, Y., & Kim, J. (2021) · Electronics · doi:10.3390/electronics10151809 · Cited by 23
FPGA-based Acceleration for Convolutional Neural Networks on PYNQ-Z2
Huynh, T. V. (2022) · International Journal of Computing and Digital Systems · doi:10.12785/ijcds/110136 · Cited by 25
“These accuracies are comparable with the accuracies of other related models reported in [25] for handwritten digit recognition applications.”
ResultNazari, S., Keyanfar, A., & Hulle, M. M. V. (2022) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2022.105430 · Cited by 24
Hasan, T., Rahim, M. A., Shin, J., Nishimura, S., & Hossain, M. N. (2024) · IEEE Access · doi:10.1109/access.2024.3352070 · Cited by 22
“Baldominos et al. [25] introduced the use of CNNs in identifying handwritten characters, focusing on the difference between previous work that used data augmentation and the original dataset.”
MethodUsing Deep Learning to Demodulate Transmissions in Molecular Communication
Bartunik, M., Keszöcze, O., Schiller, B., & Kirchner, J. (2022) · International Symposium on Medical Information and Communication Technology · doi:10.1109/ismict56646.2022.9828263 · Cited by 23
Urdu handwritten text recognition: a survey
Husnain, M., Missen, M. M. S., Mumtaz, S., Coustaty, M., Luqman, M. M., & Ogier, J.-M. (2020) · IET Image Processing · doi:10.1049/iet-ipr.2019.0401 · Cited by 21
“Some of these issues are differences in writing style (even from the same author), image degradation due to cursive nature of the script, poor quality or illegible handwriting etc. [5, 6, 13].”
BackgroundHuang, C., Cui, J., Fu, Y., Huang, D., Zhao, M., & Li, L. (2023) · Neural Networks · doi:10.1016/j.neunet.2023.03.013 · Cited by 21
Devanagari Character Recognition: A Comprehensive Literature Review
Arora, S., Malik, L., Goyal, S., Bhattacharjee, D., Nasipuri, M., & Krejcar, O. (2024) · IEEE Access · doi:10.1109/access.2024.3520248 · Cited by 30
“One of the most advanced machine learning methods, deep learning, was first presented in 1998 for character identification using the MNIST database [118].”
Selvaraj, J., & Jayanthy, A. K. (2024) · International Journal of Imaging Systems and Technology · doi:10.1002/ima.23034 · Cited by 20
Lee, A., Yu, H., & Min, G. (2024) · Journal of Cultural Heritage · doi:10.1016/j.culher.2024.02.001 · Cited by 21
Survey on Segmentation and Recognition of Handwritten Arabic Script
Ali, A. A. A., & Suresha, M. (2020) · SN Computer Science · doi:10.1007/s42979-020-00187-y · Cited by 20
Handwritten Recognition: A survey
AL-Taee, M. M., Neji, S. B. H., & Frikha, M. (2020) · IEEE International Conference on Image Processing, Applications and Systems (IPAS) · doi:10.1109/ipas50080.2020.9334936 · Cited by 20
“The CNNs represent a kind of neural network that is capable of extracting pertinent features of the multidimensional inputs data automatically and turn them into a very interesting alternative for solving problems within the field of computer vision [4].”
BackgroundAmiri, M., Jafari, A. H., Makkiabadi, B., Nazari, S., & Hulle, M. M. V. (2022) · Information Sciences · doi:10.1016/j.ins.2022.11.162 · Cited by 20
Azgar, A., Nazir, M. I., Akter, A., Hossain, M. S., Wadud, M. A. H., & Islam, M. R. (2024) · Lecture notes in networks and systems · doi:10.1007/978-981-97-3562-4_44 · Cited by 19
Differential privacy protection over deep learning: An investigation of its impacted factors
Lin, Y., Bao, L.-Y., Li, Z.-M., Si, S.-Z., & Chu, C. (2020) · Computers & Security · doi:10.1016/j.cose.2020.102061 · Cited by 21
Convolutional neural network-based ensemble methods to recognize Bangla handwritten character
Shibly, M. M. A., Tisha, T. A., Tani, T. A., & Ripon, S. (2021) · PeerJ Computer Science · doi:10.7717/peerj-cs.565 · Cited by 20
“For English handwritten digit and character recognition tasks, CNN-based architectures have yielded better performance than other techniques ( Baldominos, Saez & Isasi, 2019 ; Ranzato et al., 2007 ; Cire ş an et al., 2011 ), and so on.”
MethodDeep Learning for Handwriting Text Recognition: Existing Approaches and Challenges
Teslya, N., & Mohammed, S. (2022) · Conference of Open Innovations Association (FRUCT) · doi:10.23919/fruct54823.2022.9770912 · Cited by 18
Afkari-Fahandari, A., Shabaninia, E., Asadi-Zeydabadi, F., & Nezamabadi–pour, H. (2025) · ACM Computing Surveys · doi:10.1145/3771273 · Cited by 22
“As depicted in the table, certain papers delve into the exploration of both conventional and deep-learning-based approaches [11, 16, 122, 132, 154].”
A Novel Computational Framework for Precision Diagnosis and Subtype Discovery of Plant With Lesion
Xia, F., Xie, X., Wang, Z., Jin, S., Yan, K., & Ji, Z. (2022) · Frontiers in Plant Science · doi:10.3389/fpls.2021.789630 · Cited by 17
“The MNIST database (Modified National Institute of Standards and Technology database) (Baldominos et al., 2019), a large database of handwritten digits, was used for data collection, which is not only used for training various image processing systems but also for testing machine-learning algorithms…”
MethodEfficient construction and convergence analysis of sparse convolutional neural networks
Zhao, S., Fan, Q., Dong, Q., Xing, Z., Yang, X., & He, X. (2024) · Neurocomputing · doi:10.1016/j.neucom.2024.128032 · Cited by 16
CArDIS: A Swedish Historical Handwritten Character and Word Dataset
Yavariabdi, A., Kusetoğulları, H., Çelik, T., Thummanapally, S., Rijwan, S., & Hall, J. (2022) · IEEE Access · doi:10.1109/access.2022.3175197 · Cited by 16
Building Better Datasets: Seven Recommendations for Responsible Design from Dataset CreatorsPreprint
Orr, W., & Crawford, K. (2024) · arXiv (Cornell University) · doi:10.48550/arxiv.2409.00252 · Cited by 28
“Newer, superior, or more challenging datasets may be released (e.g., Yang et al., 2020), and models may achieve near-perfect prediction scores, rendering the dataset "solved" (e.g., Baldominos et al., 2019).”
BackgroundMNSIT Handwritten Digit Recognition Using Machine Learning Classification Algorithms
Anoop, R., Kiran, N., Raj, B. P., Kodipalli, A., Rao, T., & Rohini, B. R. (2023) · World Conference on Communication & Computing (WCONF) · doi:10.1109/wconf58270.2023.10234977 · Cited by 14
“There are various papers and articles being published about this topic these days [8][9][10][11][12][13].”
BackgroundAbosamra, G., & Oqaibi, H. M. (2021) · Computers, materials & continua/Computers, materials & continua (Print) · doi:10.32604/cmc.2021.015318 · Cited by 15
“Works done on the National Institute of Standards and Technology dataset (NIST) Special Database 19 (from which EMNIST was obtained) can be found in [3].”
MethodAlbayati, A. Q., Altaie, S. A. J., Al-Obaydy, W. N. I., & Alkhalid, F. F. (2023) · IAES International Journal of Artificial Intelligence · doi:10.11591/ijai.v13.i1.pp563-571 · Cited by 14
“To enlarge the images into a 28 × 28 pixels box and to ensure that the center of the digit’s mass matched its center, blank padding was finally applied [33].”
MethodAnalog Spiking Neural Network Synthesis for the MNIST
Soupizet, T., Jouni, Z., Wang, S., Benlarbi‐Delaï, A., & Ferreira, P. M. (2023) · Journal of Integrated Circuits and Systems · doi:10.29292/jics.v18i1.663 · Cited by 13
“Convolutional neural networks (CNNs) are the most well known and widely used techniques for the MNIST [15].”
BackgroundMethodInfluentialComparative Analysis of Handwritten Text Recognition using CNN and SVM
Parameshachari, B. D., Ashok, A., & Reddy, H. (2023) · International Conference on Distributed Computing and Electrical Circuits and Electronics (ICDCECE) · doi:10.1109/icdcece57866.2023.10150890 · Cited by 13
“The most common and organized method of recording facts and information is handwriting [6].”
MethodIzotov, Y. A., Velichko, A. A., Ivshin, А. А., & Novitskiy, R. E. (2021) · IOP Conference Series Materials Science and Engineering · doi:10.1088/1757-899x/1155/1/012056 · Cited by 13
Deep Learning-Based Optimization of Visual–Auditory Sensory Substitution
Kim, M., Park, Y., Moon, K., & Jeong, C. Y. (2023) · IEEE Access · doi:10.1109/access.2023.3243641 · Cited by 12
“Therefore, we used the MNIST handwritten digit dataset [31], which is publicly available, as the visual data for our experiment.”
MethodHumayun, M., Siddiqi, R., Uddin, M., Kandhro, I. A., Abdelhaq, M., & Alsaqour, R. A. (2024) · Neural Computing and Applications · doi:10.1007/s00521-024-10206-1 · Cited by 19
A Comprehensive Overview of Handwritten Recognition Techniques: A Survey
Ali, S. H., & Abdulrazzaq, M. B. (2023) · Journal of Computer Science · doi:10.3844/jcssp.2023.569.587 · Cited by 12
“…one of the most recognized and widely used methods in deep learning which represents a type of neural network capable of mechanically extracting features from multidimensional input data to solve difficulties in the computer's field of view, this was explained in the studies by Baldominos et al .”
MethodBahrami, M. K., & Nazari, S. (2024) · Scientific Reports · doi:10.1038/s41598-024-54043-7 · Cited by 12
“Notably, achieving high accuracy on EMNIST and CIFAR10 using spiking pattern recognition networks remains a formidable challenge 11,12 , marking this work as state-of-the-art.”
BackgroundImage Feature Learning with Genetic Programming
Ruberto, S., Terragni, V., & Moore, J. H. (2020) · Lecture notes in computer science · doi:10.1007/978-3-030-58115-2_5 · Cited by 12
Koh, R. G. L., Dilek, B., Ye, G., Selver, A., & Kumbhare, D. A. (2023) · Ultrasound in Medicine & Biology · doi:10.1016/j.ultrasmedbio.2023.06.019 · Cited by 11
A Synaptic Pruning-Based Spiking Neural Network for Hand-Written Digits Classification
Faghihi, F., Alashwal, H., & Moustafa, A. A. (2022) · Frontiers in Artificial Intelligence · doi:10.3389/frai.2022.680165 · Cited by 9
“%) on this dataset (Baldominos et al., 2019; Patil, 2020).”
BackgroundHierarchical Classification on the MNIST Dataset Using Truncated SVD and Kernel Density Estimation
Plesovskaya, E. P., & Ivanov, S. (2022) · Procedia Computer Science · doi:10.1016/j.procs.2022.11.021 · Cited by 11
Dysgraphia detection based on convolutional neural networks and child-robot interaction
Gouraguine, S., Riad, M., Qbadou, M., & Mansouri, K. (2023) · International Journal of Power Electronics and Drive Systems/International Journal of Electrical and Computer Engineering · doi:10.11591/ijece.v13i3.pp2999-3009 · Cited by 11
Integration AI Techniques in Low-Resource Language: The Case of Kazakh Language
Asmaganbetova, K., Auyeskhan, U., Zarina, N., Alen, S., & Yeskazina, M. (2024) · IEEE AITU: Digital Generation · doi:10.1109/ieeeconf61558.2024.10585350 · Cited by 13
“[3, 4, 6].”
Arsa, D. M. S., Putri, G. A. A., Zen, R. A. M., & Bressan, S. (2020) · International Conference on Knowledge and Systems Engineering/Proceedings · doi:10.1109/kse50997.2020.9287584 · Cited by 9
Recognizing intertwined patterns using a network of spiking pattern recognition platforms
Amiri, M., Jafari, A. H., Makkiabadi, B., & Nazari, S. (2022) · Scientific Reports · doi:10.1038/s41598-022-23320-8 · Cited by 9
Farhaoui, O., Fethi, M. R., Zeroual, I., & Allaoui, A. E. (2023) · Data & Metadata · doi:10.56294/dm2023176 · Cited by 9
Novel Adaptive Binarization Method for Degraded Document Images
Abdullah, S. N. H. S., Ismail, S. M., Hasan, M. K., & Shivakumara, P. (2021) · Computers, materials & continua/Computers, materials & continua (Print) · doi:10.32604/cmc.2021.014610 · Cited by 9
“The literature on document analysis revealed that optical character recognition (OCR) is a successful system in the area of image processing and pattern recognition owing to its capability to achieve more than 90% recognition for scanned plain background images [2,3].”
Rehman, M. Z., Nawi, N. M., Arshad, M., & Khan, A. (2021) · Electronics · doi:10.3390/electronics10202508 · Cited by 9
“Developed by Germany and later on used extensively in the US, OCR utilizes computer vision, which enables a machine to recognize different characters of any language [2].”
BackgroundNL-CNN: A Resources-Constrained Deep Learning Model based on Nonlinear Convolution
Dogaru, R., & Dogaru, I. (2021) · International Symposium on Advanced Topics in Electrical Engineering · doi:10.1109/atee52255.2021.9425248 · Cited by 9
“com/post/3-small-but-powerful-cnn/ [11] A.”
Optical Character Recognition and Text to Speech Generation System using Machine Learning
Padmavathi, P., Mahadas, B. B., Kalluri, S. S., Devarapu, P., & Bandi, S. L. (2023) · International Conference on Applied Artificial Intelligence and Computing (ICAAIC) · doi:10.1109/icaaic56838.2023.10140864 · Cited by 9
Analysis and Validation of Cross-Modal Generative Adversarial Network for Sensory Substitution
Kim, M., Park, Y., Moon, K., & Jeong, C. Y. (2021) · International Journal of Environmental Research and Public Health · doi:10.3390/ijerph18126216 · Cited by 8
“We used the MNIST handwritten digits dataset [47] as the input image for our experiments because it has been used as a benchmark for most previous approaches.”
MethodA Progressive Approach to Arabic Character Recognition Using a Modified Freeman Chain Code Algorithm
Fethi, M. R., Farhaoui, O., Zeroual, I., & Allaoui, A. E. (2023) · Data & Metadata · doi:10.56294/dm2023178 · Cited by 8
Efat, A. H., Hasan, S. M. M., & Zibran, M. F. (2025) · IEEE International Conference on Computing and Machine Intelligence (ICMI) · doi:10.1109/icmi65310.2025.11141266 · Cited by 8
“Baldominos et al. [10] optimized data augmentation techniques, achieving a 0.23% test error rate on the MNIST database using a CNN.”
InfluentialMitigating Group Bias in Federated Learning for Heterogeneous Devices
Selialia, K., Chandio, Y., & Anwar, F. M. (2024) · ACM Conference on Fairness, Accountability, and Transparency (FAccT) · doi:10.1145/3630106.3658954 · Cited by 15
“In this section, we first formally define our problem and then present our approach to improving the worst group performance while maintaining the best group performance.”
BackgroundEntropic associative memory for manuscript symbols
Morales, R., Hernández, N., Cruz, R., Cruz, V. D., & Pineda, L. A. (2022) · PLoS ONE · doi:10.1371/journal.pone.0272386 · Cited by 9
“We also reviewed the survey of systems using MNIST and EMNIST for handwritten character recognition [33], but associative memories are not even mentioned, and we were unable to find a single associative memory using the EMNIST corpus for direct comparisons.”
A Novel Unsupervised Spatial–Temporal Learning Mechanism in a Bio-inspired Spiking Neural Network
Amiri, M., Jafari, A. H., Makkiabadi, B., & Nazari, S. (2022) · Cognitive Computation · doi:10.1007/s12559-022-10097-1 · Cited by 8
“In addition to the results reported in the above tables, many related studies are presented in [74].”
ResultConvolutional Neural Networks with Hebbian-Based Rules in Online Transfer Learning
Aguilar-Canto, F. J. (2020) · Lecture notes in computer science · doi:10.1007/978-3-030-60884-2_3 · Cited by 11
Quantitative Analysis of Deep CNNs for Multilingual Handwritten Digit Recognition
Haque, M. R., Azam, M. G., Milon, S. M., Hossain, M. S., Molla, M. A.-A., & Uddin, M. S. (2020) · Advances in intelligent systems and computing · doi:10.1007/978-981-33-4673-4_2 · Cited by 8
Urazoe, K., Kuroki, N., Hirose, T., & Numa, M. (2020) · IEEJ Transactions on Electrical and Electronic Engineering · doi:10.1002/tee.23278 · Cited by 8
“We use a handwritten digit dataset called MNIST with 60 000 training images and 10 000 testing images [1,7].”
BackgroundMethodAutomating selective area electron diffraction phase identification using machine learning
Mika, M., Tomczak, N., Finney, C., Carter, J. L. W., & Aitkaliyeva, A. (2024) · Journal of Materiomics · doi:10.1016/j.jmat.2023.12.010 · Cited by 8
Amiri, M., & Nazari, S. (2024) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2024.109255 · Cited by 7
Fu, M., Wang, Q., Wang, J., Sun, L., Ma, Z., Zhang, C., Guan, W., Liu, Q., Wang, D., & Li, W. (2023) · Applied Sciences · doi:10.3390/app13021066 · Cited by 7
Papchenko, O., Kuzikov, B., & Шовкопляс, О. (2023) · RADIOELECTRONIC AND COMPUTER SYSTEMS · doi:10.32620/reks.2023.4.06 · Cited by 7
Singh, S. K., & Chaturvedi, A. (2024) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2024.109225 · Cited by 11
Implementation of generative adversarial networks in HPCC systems using GNN bundle
Karthik, A., Shetty, J., Shobha, G., & Dev, R. (2021) · IAES International Journal of Artificial Intelligence · doi:10.11591/ijai.v10.i2.pp374-381 · Cited by 7
“Simple generative adversarial networks It can be noticed in Figure 5 that simple GAN produces results very similar to the MNIST dataset [26], [27], but there are noisy pixels around them due it being a purely dense network relying on the density of each tensor.”
ResultFaouci, S., Gaceb, D., & Haddad, M. (2021) · Lecture notes on data engineering and communications technologies · doi:10.1007/978-3-030-70713-2_100 · Cited by 7
Romero‐Zaliz, R., Cantudo, A., Pérez, E., Jiménez-Molinos, F., Wenger, C., & Roldán, J. B. (2021) · Electronics · doi:10.3390/electronics10243141 · Cited by 8
Online Handwriting Based Adult and Child Classification using Machine Learning Techniques
Shin, J., Hasan, M. A. M., Maniruzzaman, M., Megumi, A., Suzuki, A., & Yasumura, A. (2022) · IEEE Eurasian Conference on Educational Innovation (ECEI) · doi:10.1109/ecei53102.2022.9829467 · Cited by 8
Overhead-MNIST: Machine Learning Baselines for Image ClassificationPreprint
Larsen, E., Noever, D. A., MacVittie, K., & Lilly, J. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2107.00436 · Cited by 7
“One popular dataset used for this purpose is the MNIST collection of handwritten numbers [7].”
BackgroundFree-space optical spiking neural network
Ahmadi, R., Ahmadnejad, A., & Koohi, S. (2024) · PLoS ONE · doi:10.1371/journal.pone.0313547 · Cited by 6
The windowed two-dimensional graph fractional Fourier transform
Gan, Y., Chen, J.-Y., & Li, B. (2025) · Digital Signal Processing · doi:10.1016/j.dsp.2025.105191 · Cited by 11
Sánchez‐Nielsen, E., & Martín-Herrera, I. (2025) · IEEE Access · doi:10.1109/access.2025.3620686 · Cited by 6
“The EMNIST dataset was selected due to its suitability as a benchmark for handwritten character recognition tasks [31].”
Simply Fine-Tuned Deep Learning-Based Classification for Breast Cancer with Mammograms
Mudeng, V., Jeong, J.-W., & Choe, S. (2022) · Computers, materials & continua/Computers, materials & continua (Print) · doi:10.32604/cmc.2022.031046 · Cited by 9
“Nowadays, LeCun et al.’s study is famous because it provides a publicly accessible dataset of handwritten digits, called the modified national institute of standards and technology (MNIST) [8] containing 60000 training and 10000 test datasets.”
KurdSet Handwritten Digits Recognition Based on Different Convolutional Neural Networks Models
Ali, S. H., & Abdulrazzaq, M. B. (2024) · TEM Journal · doi:10.18421/tem131-23 · Cited by 6
End-to-End Analysis for Text Detection and Recognition in Natural Scene Images
Alnefaie, A., Gupta, D., Bhuyan, M. H., Razzak, I., Gupta, P. K., & Prasad, M. (2020) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn48605.2020.9207001 · Cited by 6
“[36] review work contributions evaluated on the EMNIST dataset for handwritten text recognition.”
BackgroundConsistent Lock-free Parallel Stochastic Gradient Descent for Fast and Stable Convergence
Bäckström, K., Walulya, I., Papatriantafilou, M., & Tsigas, P. (2021) · IEEE International Parallel and Distributed Processing Symposium (IPDPS) · doi:10.1109/ipdps49936.2021.00051 · Cited by 6
An Intelligent Handwritten Digits and Characters Recognition System
Mokhtar, A. A., El-Sharkawy, A. M., Mahmoud, S. S., Samy, H. W., AbdelMawla, B., Moussa, K., & Darweesh, M. S. (2022) · International Conference on Innovation and Intelligence for Informatics, Computing, and Technologies (3ICT) · doi:10.1109/3ict56508.2022.9990669 · Cited by 6
Metadata
- Authors
- Baldominos, A., Saez, Y., & Isasi, P.
- Affiliations
- Universidad Carlos III de Madrid (ES)
- Computer Science Department, Universidad Carlos III of Madrid, 28911 Leganés, Madrid, Spain
- Corresponding author
- Baldominos, Alejandro
- Venue
- Applied Sciences
- Publisher
- MDPI AG
- Published
- 2019-08-04
- Volume
- 9
- Issue
- 15
- Pages
- 3169
- ISSN
- 2076-3417 (ISSN-L)
- OpenAlex
- W2966275995
- Cited by
- 279
- References
- 76
- Citations by year
- 2026: 20 · 2025: 39 · 2024: 64 · 2023: 54 · 2022: 41 · 2021: 36 · 2020: 24 · 2019: 1
- FWCI
- 11.41
- Type
- journal article
- Language
- en
- Open access
- gold
- dx.doi.org
- Licence
- Locations
- Applied Sciences (publishedVersion, cc-by)
- e-Archivo (Carlos III University of Madrid) (submittedVersion, cc-by)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion)
- Funders
- Ministerio de Educación, Cultura y Deporte (FPU13/03917)
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Handwritten Text Recognition Techniques
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Vehicle License Plate Recognition
- Keywords
- handwritten character recognition, handwritten digit recognition, MNIST, EMNIST, convolutional neural networks, data augmentation, letter recognition
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.


