Classification of heterogeneous Malayalam documents based on structural features using deep learning models

International Journal of Electrical and Computer Engineering (IJECE)
Vol. 13, No. 1, February 2023, pp. 894~901
ISSN: 2088-8708, DOI: 10.11591/ijece.v13i1.pp894-901  894
Journal homepage: http://ijece.iaescore.com
Classification of heterogeneous Malayalam documents based on
structural features using deep learning models
Bipin Nair Balakrishnan Jayakumari, Amel Thomas Kavana
Department of Computer Science, Amrita School of Arts and Sciences, Mysuru Campus, Amrita Vishwa Vidyapeetham,
Tamil Nadu, India
Article Info ABSTRACT
Article history:
Received Mar 21, 2022
Revised Jul 17, 2022
Accepted Aug 21, 2022
The proposed work gives a comparative study on performance of various
pretrained deep learning models for classifying Malayalam documents such
as agreement documents, notebook images, and palm leaves. The documents
are classified based on their visual and structural features. The dataset was
manually collected from different sources. The method of research proceeds
with preprocessing, feature extraction, and classification. The proposed work
deals with three fine-tuned deep learning models such as visual geometry
group-16 (VGG-16), convolutional neural network (CNN) and AlexNet. The
models attained high accuracies of 99.7%, 96%, and 95%, respectively.
Among the three models, the fine-tuned VGG-16 model was found to
perform better attaining a very high accuracy on the dataset. As a future
work, methods to classify the documents based on content as well as spectral
features can be developed.
Keywords:
Classification
Deep learning
Documents
AlexNet
Preprocessing
This is an open access article under the CC BY-SA license.
Corresponding Author:
Bipin Nair Balakrishnan Jayakumari
Department of Computer Science, Amrita School of Arts and Sciences
Amrita Vishwa Vidyapeetham, Tamil Nadu 641112, India
Email: bipin.bj.nair@gmail.com
1. INTRODUCTION
Ancient documents reveal the history of the people, nation as well as tradition. The preservation and
segregation of these documents is a tedious process. The documents get degraded over time due to various
natural factors like aging, environmental factors, and accidental errors [1]. Digitization of documents is an
effective method for the preservation as well as classifying the documents.
Document classification will throw a new light on the new era of digitization and categorization. As
the need for digitization increases, the requirement for the classification of these digitized documents comes
into play. Document classification is the process of classifying or grouping documents into various categories
based on the structural features. Document segregation and storage is an important step in information
management and retrieval [2]. As each document belongs to different categories, classifying the documents
manually would consume more time and effort. A deep learning-based method is proposed to check which
classifier obtains high accuracy for the classification of ancient Malayalam documents based on structural
features. The documents are classified into three categories namely palm leaves, agreement copies, and
notebook images.
Mushtaq et al. [1] investigated a deep learning-based convolutional neural network (CNN) model
for spectral image classification on datasets which consist of 10, 10 and 50 classes respectively have got a
good accuracy of 99.04%, 99.49%, and 97.57% were reported respectively for each dataset. In [2], [3] a deep
model based on CNN was adopted for classifying documents that was experimented on various datasets
containing over 5,200 documents got remarkable accuracies compared to the traditional models. Quan
et al. [4] discussed a multispectral fusion-based classification on radar image and sentinel-2A datasets with

Int J Elec & Comp Eng ISSN: 2088-8708 
Classification of heterogeneous Malayalam documents based … (Bipin Nair Balakrishnan Jayakumari)
895
adequate accuracy. Prieto et al. [5] proposed a deep model probabilistic indexing to perform classification on
un-transcribed manuscript images from Spanish Archivo General de Indias that could achieve an accuracy of
70% on 3 different classes. Nasir et al. [6] investigated a deep model AlexNet and visual geometry group-19
(VGG-19) on Tobacco-3842 dataset consisting of 10 different image classes and a dataset containing 15
different classes achieved an exceptional accuracy rate of 93.1%. Bakkali et al. [7] discussed a deep
classification model Nasnet on Tobacco-3842 to give an overall accuracy of 99.7% over 10 different classes.
A classification graph based neural network model was investigated by Mandivarapu et al. [8] on insurance
and Tobacco 3842 dataset which give an adequate accuracy of 91% for 11 classes and 77.5% for 10 different
classes. Chen et al. [9] experimented a decomposition-based hyperspectral classification approach on the Sar
images dataset which achieved a remarkable accuracy of 99.7%. CNN was found to be effective for
classifying hyperspectral images [10], [11] resulting in faster processing and higher classification accuracies
over various hyperspectral image datasets.
A novel model was investigated by Gayathri and Kannan [12] on a dataset containing over
3,000 images to classify Ayurvedic documents which gave promising results. Jimenez et al. [13] proposed a
deep neural network model to classify Covid-19 literature on the LitCovid and Covid-19 open research
dataset (CORD-19) datasets containing 8 classes each showed better accuracies than the existing traditional
methods. A deep learning-based Squeeze net model on Tobacco-3842 dataset consisting of 10 different
classes was discussed by Hassanpour et al. [14] to classify documents based on visual features with an
accuracy of 77% was obtained. Kanchi et al. [15] discussed a deep multi model-based approach to classify
documents on datasets containing 16 and 10 classes respectively. The proposed approach obtained an
accuracy of 90.3%. A deep active learning-based approach was investigated by Hemmer et al. [16] for the
classification of images that attained an accuracy of 90% on the Modified National Institute of Standards and
Technology (MNIST) and Cifar-10 datasets. Indraswari et al. [17] proposed a mobile net-based classification
of melanoma images which achieved an accuracy of 85% over four different datasets containing images
belonging to two classes. Ahmed et al. [18] investigated a deep neural net model with attention mechanism
for Bangla document classification on a manually collected dataset. The proposed model obtained an
accuracy rate of 86.56% was obtained over 13 document classes. Pan et al. [19] experimented with an
ontology-driven approach to classify scientific literature that achieved a score of 95% on DBLP dataset.
Jiang et al. [20] proposed three various deep models for technical documents for the classification that could
yield a decent accuracy rate of 77.9% over 50 distinct classes. A deep learning-based adaptive multiscale
segmentation method was proposed by Zhao et al. [21] on Indian Pines, Salinas Scene and University of
Pavia datasets containing 16.9 and 15 classes respectively on which accuracies of 94.312%, 99.217%, and
92.693% were obtained. A deep learning-based hybrid machine learning based model was developed by
Swetanisha et al. [22] for classifying multispectral images on Landsat-8 dataset containing 7 different classes
of satellite images could attain decent accuracy scores. In [23], [24] a combined approach of deep learning
and machine learning models were used to perform multiclass classification of documents which could yield
exceptionally high accuracy values. Deep neural network-based models were proposed in [25]–[27] for
classification of hyperspectral images on the Indian Pines and University of Houston and Salinas Seas
datasets containing over 16 image classes. The proposed models could attain a high accuracy rate varying
between 90-99%. Jayakumari and Nair [28] proposed a deep learning based ResNet model to perform
binarization of ancient horoscopic palm leaf images which attained a very high accuracy of 95.38% on a
manually collected dataset consisting of ancient horoscopic palm leaves. Deep learning is definitely the pick
of the bunch when the problem requires processing of huge and unstructured deep model data processing
[29]. The further sections of paper are described below as methods, results and discussion, and conclusion.
2. METHOD
The proposed method classifies ancient documents such as agreement copies, palm leaf manuscripts,
and notebook images based on their structural features. The methodology has three approaches using three
different deep learning models such as CNN, fine-tuned VGG-16 and modified AlexNet along with various
enhancement methods for classifying the documents. Each model is evaluated on the basis of the accuracy
obtained over the dataset. The model that performs better on the dataset is identified from the evaluation results.
2.1. Data collection
The proposed work is to classify Malayalam documents belonging to three different categories of
documents such as agreement copies, palm leaves, and notebook images. The datasets used for our proposed
work are manually collected from people as well as from online repositories. The datasets obtained contained
degraded documents which made them difficult to be classified because of the same pattern of Malayalam
characters present in all the documents.

 ISSN: 2088-8708
Int J Elec & Comp Eng, Vol. 13, No. 1, February 2023: 894-901
896
Table 1 displays the details about the datasets, their sources, and the number of samples collected.
The agreement copies were manually collected from various Taluk offices as well as from the internet. The
palm leaf data set is collected from various people in Kerala as well as from Varikkasheri Mana, Palakkad,
Kerala; and notebook images [30] are collected from various schools and colleges of Kottayam Kerala. For
the research, 1,500 samples of each category of documents were collected.
Table 1. Dataset collection details
S.no Source Dataset Count
1 Internet, Taluk offices Agreement copies 1,500
2 Kerala Manas Palm leaves 1,500
3 Universities, Schools, Online repositories Notebook images 1,500
2.2. Preprocessing
In the initial approach using CNN, the input image is resized to 224×224 and is converted into
grayscale format. The grayscale image then undergoes Otsu thresholding for image enhancement.
Thresholding is done to binarize the input image based on its pixel intensity values. The Otsu enhancement
process uses (1).
σ2(x)=ωbg(x)σ2bg(x)+ωfg(x)σ2fg(x) (1)
In (1), ωbg(x) and ωfg(x) are the probability of the number of pixels of each class at a threshold value
of X. σ2
represents the color value variance. In the second approach using VGG-16, the enhancement process
is done by normalizing the RGB values for each pixel of the input document image. Here, the mean pixel
value is reduced from each pixel in this process. The image is normalized and resized to 224×224. The
normalization enhancement process uses (2).
𝑦𝑖 =
𝑠𝑖−𝑚𝑖𝑛(𝑠)
𝑚𝑎𝑥(𝑠)−𝑚𝑖𝑛(𝑠)
(2)
In (2), s is the input data that ranges from s1 to sn, and yi becomes the ith
normalized data. In the
approach using AlexNet, the input image is initially normalized by rescaling. The image is then resized to
227×227 as it is the standard input size for AlexNet architecture. In third approach also uses the same
pre-processing method as normalization.
2.3. Classification using CNN
In Figure 1, after pre-processing the image is taken as the input for the CNN model. The model
consists of three stacks of convolutional and MaxPooling layers, a flattening layer, and three dense layers.
The input image of size 224×224 is passed through a convolutional layer having 32 filters of size 3×3 with
ReLU activation function from which it is passed to a max pooling layer of filter size 2×2. The image is then
inputted into the next convolution layer with filter having 64 filters of size 3×3. It is then passed to the
MaxPooling layer of filter size 2×2 that follows. The image is forwarded to the next convolutional layer
having number of filters 64 of size 3×3. It is then inputted to the next MaxPooling layer of filter size 3×3 and
a dropout layer that follows. The output image is flattened and fed into the hidden dense layer with 128 filters
and ReLU activation function. The layer that follows is the output layer trained for three classes.
Figure 1. Classification using CNN

897
2.4. Classification using fine-tuned VGG-16
In Figure 2, the enhanced image is passed through the VGG-16 network which is a stack of two
convolutional layers having 64 filters of size 3×3, from which it is passed to max pooling layer of filter size
2×2. The image size is reduced to 112×112×64 after max pooling. The image is then passed to the next stack
of two convolution layers and a max pooling layer where the same process is repeated and as a result, an
output image of size 56×56×128 is obtained. This is followed by the next stack containing three convolution
layers of kernel size 256 which makes the output size 28×28×256. The next two consecutive stacks again
contain three convolution layers and 512 filters each. After the image passes through the two stacks, the
output will be of size 7×7×512. The obtained output is flattened and passed to a stack of three fully
connected layers. The model is fine-14 by replacing the final fully connected dense layer which serves as the
output layer.
Figure 2. Classification using fine-tuned VGG-16
Figure 3 depicts pseudo code of the VGG-16 model. The processing method displays how the image
is processed by the VGG-16 model. In the figure, Iimg stands for the input image which is of size 224×224
which is fed into the VGG-16 model, Stack stands for the stack of convolutional and MaxPooling layers,
conv stands for convolutional layer, L stands for layer, Maxpool is the abbreviation for MaxPooling layer,
Flat is used to denote flattening layer and Oimg is the output image.
img
to )
Maxpool L(1 to 5)
Conv L(1 13
(224 * 224) 16
processing VGG-16()
{
Stack L (1 5)
{
}
Flat L (1 o 1)
Dense (1 to 3)
}
img
to
I VGG
O
→ −
Figure 3. VGG-16 pseudo code
2.5. Classification using modified AlexNet
The AlexNet model in Figure 4 consists of five stacks of convolutional layers activated by ReLU
activation. Each convolution layer is followed by a batch normalization layer. The input image if size
227×227 is passed to the input convolution layer with 32 filters of size 11×11 with an activation function
ReLU which is then forwarded to the batch normalization layer. The output image from the first stack of
layers is then sent to a MaxPooling layer and further forwarded to the next stack of layers. Here the number
of filters of the convolutional layer is increased to 64 and the filter size becomes 5×5. After MaxPooling, the

 ISSN: 2088-8708
898
modified image is sent to the following stack of three convolutional layers with 128, 128, and 256 filters
respectively of 3×3 size each. The image is then flattened and forwarded to a couple of dense layers each
with ReLU activation. It is then forwarded to the output layer that has been trained for classifying the three
classes. The output layer is activated by SoftMax activation.
Figure 4. Classification using modified AlexNet
3. RESULTS AND DISCUSSION
The experiment was carried out on a dataset consisting of a total of 4,500 images belonging to three
different classes. Out of which 3,600 images were used for training, 705 images for testing and 195 images
for predictions. The models were tested on the dataset and the performance of each model has been assessed
based on the performance evaluation metrics. Tables 2 to 4 show the performance evaluation metrics of each
model.
Table 2. Performance evaluation result of
classification using CNN
Epochs Accuracy Precision Recall F1 score
1 91.34 0.590 0.670 0.590
2 94.58 0.590 0.860 0.660
3 93.22 0.630 0.710 0.560
4 94.55 0.570 0.570 0.340
5 94.85 0.500 0.620 0.500
6 96.23 0.420 0.560 0.400
7 91.20 0.590 0.600 0.490
8 92.05 0.590 0.600 0.490
9 90.56 0.690 0.650 0.550
10 88.11 0.750 0.600 0.590
Table 3. Performance evaluation result of
classification using fine-tuned VGG-16
1 88.6 0.850 0. 850 0.850
2 84.8 0.900 0.900 0.900
3 81.1 0.850 0.850 0.850
4 85.6 0.900 0.900 0.900
5 91.6 0.800 0.800 0.800
6 94.2 0.850 0.850 0.850
7 99.7 0.950 0.950 0.950
8 95.5 0.900 0.900 0.900
9 93.5 0.880 0.880 0.880
10 94.8 0.840 0.840 0.840
Table 4. Performance evaluation result of classification using AlexNet
1 43.8 0.390 0.490 0.490
2 68.3 0.590 0.860 0.660
3 85.3 0.630 0.710 0.500
4 91.2 0.820 0.670 0.540
5 95.0 0.410 0.620 0.440
6 94.4 0.720 0.880 0.600
7 95.1 0.590 0.600 0.490
8 92.4 0.590 0.600 0.490
9 82.0 0.690 0.650 0.550
10 86.8 0.630 0.700 0.690
Table 2 displays the accuracy, precision, recall, and F1-score values obtained by the proposed CNN
method. The accuracy is found to be increasing gradually with the number of epochs. The maximum
accuracy is obtained at the 6th
epoch after which the accuracy tends to decrease gradually. The precision,
recall, and F1-score values varied inconsistently with each epoch.

899
From Table 3, it is observed that the VGG-16 model accuracy tends to decrease in the initial epochs.
However, from the 4th
epoch, accuracy increases gradually. The highest accuracy is obtained at the 7th
epoch
after which a decline in the accuracy values can be witnessed. The model could successfully classify the
input documents into the respective classes and achieved a very high accuracy of 99.7% on the dataset. The
values such as precision, recall, and F1-score were found to be high and were balanced as the epochs varied.
Table 4 shows the performance of the modified AlexNet on the test dataset. The model accuracy
improves as the epoch’s increases and reaches a maximum of 95.5% at the 7th
epoch. However, a decline was
observed in accuracy after the 7th
epoch. The precision, recall, and F1-score values were found to be
decreasing after the sixth epoch.
Figure 5 depicts the accuracy, loss values of the CNN, fine-tuned VGG-16 and modified AlexNet
models while training. From the accuracy loss graph of CNN, it can be inferred that the accuracy of the
model increases with the epoch and reaches a very high value near 90%. Meanwhile, the loss reduces after
the initial epoch, it finally reaches a very low value. For the VGG-16 model as the number of epoch’s
increases, the accuracy increases gradually and reaches a value in the range of 80 to 95% whereas the loss
reduces to a value in the range 0.4% to 0.6%. Finally, the performance of the modified AlexNet model is
depicted in the third graph. It is observed that the loss sharply declined in the model loss with the decrease
in the epochs. Meanwhile, accuracy is found to be increasing with the number of epochs and reaches a
value in the range 90 to 100%.
(a) (b) (c)
Figure 5. The training accuracy, loss graphs of (a) the CNN model, (b) fine-tuned VGG-16 model, and
(c) modified AlexNet model
Table 5 shows the misclassifications by AlexNet. Compared to CNN and VGG-16 model, it was
observed that in AlexNet notebook images were often confused with agreement images. A few palm leaf
images were also misclassified as notebook images.
The graph in Figure 6 depicts the extent of homogeneity among the datasets of the three different
classes.it clearly shows the structure wise similarity among notebook images, agreement copies as well as
palm leaves. From the graph we can conclude that notebook and agreement images have more similarity due
to which more documents are misclassified as notebook instead of agreement copies and agreements to
notebook.
Table 5. Misclassifications by AlexNet model
S.no Document Misclassification
1 Agreement Classified as notebook
3 Palmleaf Classified as agreement
4 Notebook Classified as agreement
5 Palmleaf Classified as notebook
9 Palmleaf Classified as notebook

 ISSN: 2088-8708
900
Figure 6. The homogeneity values between the three classes
4. CONCLUSION
The fine-tuned VGG-16 model was found to be achieving dominant results over the other two
models with a remarkable accuracy rate of 99.7%. The CNN model achieved a good accuracy score of 96%
whereas the modified AlexNet achieved an impressive accuracy of 95%. The AlexNet was found to give
more misclassified results. The proposed approach can be used to perform document classification based on
structural and visual features. This method of automatic classification of documents can be used to replace
manual classification of documents for purposes such as document digitization, and cataloguing. The future
work is to increase the number of classes by including several different categories of documents and to
perform intra class classification based on textual as well as spectral contents.
ACKNOWLEDGEMENTS
We would like to thank the government offices and archive centers for providing sufficient dataset
for our experimentation.
REFERENCES
[1] Z. Mushtaq, S.-F. Su, and Q.-V. Tran, “Spectral images based environmental sound classification using CNN with meaningful
data augmentation,” Applied Acoustics, vol. 172, Jan. 2021, doi: 10.1016/j.apacoust.2020.107581.
[2] R. Arief, A. Benny Mutiara, T. Maulana Kusuma, and H. Hustinawaty, “Automated hierarchical classification of scanned
documents using convolutional neural network and regular expression,” International Journal of Electrical and Computer
Engineering (IJECE), vol. 12, no. 1, pp. 1018–1029, Feb. 2022, doi: 10.11591/ijece.v12i1.pp1018-1029.
[3] M. M. Rahman, R. Sadik, and A. A. Biswas, “Bangla document classification using character level deep learning,” in 2020 4th
International Symposium on Multidisciplinary Studies and Innovative Technologies (ISMSIT), Oct. 2020, pp. 1–6, doi:
10.1109/ISMSIT50672.2020.9254416.
[4] Y. Quan, Y. Tong, W. Feng, G. Dauphin, W. Huang, and M. Xing, “A novel image fusion method of multi-spectral and SAR
images for land cover Classification,” Remote Sensing, vol. 12, no. 22, Nov. 2020, doi: 10.3390/rs12223801.
[5] J. R. Prieto, V. Bosch, E. Vidal, C. Alonso, M. C. Orcero, and L. Marquez, “Textual-content-based classification of bundles of
untranscribed manuscript images,” in 2020 25th International Conference on Pattern Recognition (ICPR), Jan. 2021,
pp. 3162–3169, doi: 10.1109/ICPR48806.2021.9412688.
[6] I. M. Nasir et al., “Pearson correlation-based feature selection for document classification using balanced training,” Sensors,
vol. 20, no. 23, Nov. 2020, doi: 10.3390/s20236793.
[7] S. Bakkali, Z. Ming, M. Coustaty, and M. Rusinol, “Cross-modal deep networks for document image classification,” in 2020
IEEE International Conference on Image Processing (ICIP), Oct. 2020, pp. 2556–2560, doi: 10.1109/ICIP40778.2020.9191268.
[8] J. K. Mandivarapu, E. Bunch, Q. You, and G. Fung, “Efficient document image classification using region-based graph neural
network,” preprint arXiv:2106.13802, Jun. 2021, [Online]. Available: http://arxiv.org/abs/2106.13802.
[9] G. Chen, L. Wang, and M. M. Kamruzzaman, “Spectral classification of ecological spatial polarization SAR image based on
target decomposition algorithm and machine learning,” Neural Computing and Applications, vol. 32, no. 10, pp. 5449–5460, May
2020, doi: 10.1007/s00521-019-04624-9.
[10] A. Paul, S. Bhoumik, and N. Chaki, “SSNET: an improved deep hybrid network for hyperspectral image classification,” Neural
Computing and Applications, vol. 33, no. 5, pp. 1575–1585, Mar. 2021, doi: 10.1007/s00521-020-05069-1.
[11] A. Sha, B. Wang, X. Wu, and L. Zhang, “Semisupervised classification for hyperspectral images using graph attention networks,”
IEEE Geoscience and Remote Sensing Letters, vol. 18, no. 1, pp. 157–161, Jan. 2021, doi: 10.1109/LGRS.2020.2966239.
[12] M. Gayathri and R. J. Kannan, “Ontology based concept extraction and classification of ayurvedic documents,” Procedia
Computer Science, vol. 172, pp. 511–516, 2020, doi: 10.1016/j.procs.2020.05.061.
[13] B. Jimenez Gutierrez, J. Zeng, D. Zhang, P. Zhang, and Y. Su, “Document classification for COVID-19 literature,” in Findings of
the Association for Computational Linguistics: EMNLP 2020, 2020, pp. 3715–3722, doi: 10.18653/v1/2020.findings-emnlp.332.
[14] M. Hassanpour and H. Malek, “Learning document image features with squeeze net convolutional neural network,” International
Journal of Engineering, vol. 33, no. 7, Jul. 2020, doi: 10.5829/ije.2020.33.07a.05.
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
1
47
93
139
185
231
277
323
369
415
461
507
553
599
645
691
737
783
829
875
921
967
1013
1059
1105
1151
1197
1243
1289
1335
1381
1427
1473
Feature
Values
Dataset Count
Homogeneity of classes
Palmleaf
Notebook
Agreement

901
[15] S. Kanchi, A. Pagani, H. Mokayed, M. Liwicki, D. Stricker, and M. Z. Afzal, “EmmDoc classifier: Efficient multimodal
document image classifier for scarce data,” Applied Sciences, vol. 12, no. 3, Jan. 2022, doi: 10.3390/app12031457.
[16] P. Hemmer, N. Kühl, and J. Schöffer, “DEAL: Deep evidential active learning for image classification,” in 2020 19th IEEE
International Conference on Machine Learning and Applications (ICMLA), 2022, pp. 171–192, doi: 10.1007/978-981-16-3357-
7_7.
[17] R. Indraswari, R. Rokhana, and W. Herulambang, “Melanoma image classification based on MobileNetV2 network,” Procedia
Computer Science, vol. 197, pp. 198–207, 2022, doi: 10.1016/j.procs.2021.12.132.
[18] M. Ahmed, P. Chakraborty, and T. Choudhury, “Bangla document categorization using deep RNN model with attention
mechanism,” in Cyber Intelligence and Information Retrieval, 2022, pp. 137–147.
[19] Z. Pan, P. Soong, and S. Rafatirad, “Ontology-driven scientific literature classification using clustering and self-supervised
learning,” EasyChair Preprint, pp. 1–19, 2022.
[20] S. Jiang, J. Hu, C. L. Magee, and J. Luo, “Deep learning for technical document classification,” IEEE Transactions on
Engineering Management, pp. 1–17, 2022, doi: 10.1109/TEM.2022.3152216.
[21] C. Zhao, B. Qin, S. Feng, and W. Zhu, “Multiple superpixel graphs learning based on adaptive multiscale segmentation for
hyperspectral image classification,” Remote Sensing, vol. 14, no. 3, Jan. 2022, doi: 10.3390/rs14030681.
[22] S. Swetanisha, A. R. Panda, and D. K. Behera, “Land use/land cover classification using machine learning models,” International
Journal of Electrical and Computer Engineering (IJECE), vol. 12, no. 2, pp. 2040–2046, Apr. 2022, doi:
10.11591/ijece.v12i2.pp2040-2046.
[23] C. Liu, J. Li, Q. Tang, J. Qi, and X. Zhou, “Classifying the Nunivak Island coastline using the random forest Integration of the
sentinel-2 and ICESat-2 data,” Land, vol. 11, no. 2, Feb. 2022, doi: 10.3390/land11020240.
[24] İ. Yelmen, “Multi-class document classification based on deep neural network and Word2Vec,” Journal of Aeronautics and Space
Technologies, vol. 15, no. 1, pp. 59–6, 2022.
[25] P. M. Rajegowda and P. Balamurugan, “A neural network approach to identify hyperspectral image content,” International
Journal of Electrical and Computer Engineering (IJECE), vol. 8, no. 4, pp. 2115–2125, Aug. 2018, doi:
10.11591/ijece.v8i4.pp2115-2125.
[26] C. Shi, D. Liao, T. Zhang, and L. Wang, “Hyperspectral image classification based on 3D coordination attention mechanism
network,” Remote Sensing, vol. 14, no. 3, Jan. 2022, doi: 10.3390/rs14030608.
[27] C. Ding et al., “Hyperspectral image classification promotion using clustering inspired active learning,” Remote Sensing, vol. 14,
no. 3, Jan. 2022, doi: 10.3390/rs14030596.
[28] B. N. B. Jayakumari and A. S. Nair, “Ancient horoscopic palm leaf binarization using A deep binarization model - RESNET,” in
2021 5th International Conference on Computing Methodologies and Communication (ICCMC), Apr. 2021, pp. 1524–1529, doi:
10.1109/ICCMC51019.2021.9418461.
[29] A. J. Moshayedi, A. S. Roy, A. Kolahdooz, and Y. Shuxin, “Deep learning application pros and cons over algorithm,” EAI
Endorsed Transactions on AI and Robotics, vol. 1, pp. 1–13, Feb. 2022, doi: 10.4108/airo.v1i.19.
[30] N. Sharma, “Vedic literature in Malayalam,” Ayur Malayalam Notes, 2021. https://archive.org/details/AyurMalayalamNotes
(accessed Aug. 07, 2021).
BIOGRAPHIES OF AUTHORS
Bipin Nair Balakrishnan Jayakumari is currently an assistant professor at the
Department of Computer Science, Amrita School of Arts and Sciences, Amrita
Vishwavidyapeetham, Mysuru Campus. His research interests include document image
processing, medical image processing, and image analysis. He has published more than 60
research papers in reputable national and international journals. All those research papers are
indexed in leading indexing databases such as Scopus, and SCI. He can be contacted at
bipin.bj.nair@gmail.com.
Amel Thomas Kavana is a final year student of Integrated MCA pursuing a
dissertation in the area of document image classification and deep learning at Amrita School
of Arts and Sciences, Mysuru Campus. His area of research includes document image
processing and image classification. He can be contacted at amel.t.k05@gmail.com.

Classification of heterogeneous Malayalam documents based on structural features using deep learning models

More Related Content

Similar to Classification of heterogeneous Malayalam documents based on structural features using deep learning models

More from IJECEIAES

Recently uploaded

Classification of heterogeneous Malayalam documents based on structural features using deep learning models