CNN based Optical Character Recognition and Applications

Sarika, Naragudem; Sirisala, Nageswararao; Velpuru, Muni Sekhar

doi:10.1109/icict50816.2021.9358735

Cited by 20 publications

(6 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…India is considered as one of the most multilanguage country. There are 23 language is the ancient language having the consonants and vowel count of 16 and 32 respectively [4].Also, that characters that are designed through combining the consonants and vowels are 560 'guninthalu' having 612 letters. Due to different categories this becomes complex in identifying the Hand Written Telugu language letters particularly in 'guninthalu' [5].…”

Section: Introductionmentioning

confidence: 99%

“…When learning initially starts, these characteristics may have an impact on both the network design and the training. N. Sarika et al [23], The Convolutional model is working according by first detecting Telugu text as image letter by letter, analysing the image file, and then transforming the input image into ASCII codes.The OCR technology is used to convert text contained in an image into textual form. Pre-processing, character segmentation, features extraction, and postprocessing are the three primary components of the OCR technique.This recognition model produces an accuracy with 92% and it was trained using a Telugu characters large dataset that can contain up to 1600 characters.…”

Section: Introductionmentioning

confidence: 99%

“…Finally, SVM and random forest (RF) are used to classify these features. From the research works [22][23][24][25][26][27], it is observed that Deep Learning (DL) models are implemented in recognizing the parameters. The summary of these existing works are listed under Table 1.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Unicode-driven Deep Learning Handwritten Telugu-to-English Character Recognition and Translation System

BV Subba Rao

2023

IJRITCC

View full text Add to dashboard Cite

Telugu language is considered as fourth most used language in India especially in the regions of Andhra Pradesh, Telangana, Karnataka etc. In international recognized countries also, Telugu is widely growing spoken language. This language comprises of different dependent and independent vowels, consonants and digits. In this aspect, the enhancement of Telugu Handwritten Character Recognition (HCR) has not been propagated. HCR is a neural network technique of converting a documented image to edited text one which can be used for many other applications. This reduces time and effort without starting over from the beginning every time. In this work, a Unicode based Handwritten Character Recognition(U-HCR) is developed for translating the handwritten Telugu characters into English language. With the use of Centre of Gravity (CG) in our model we can easily divide a compound character into individual character with the help of Unicode values. For training this model, we have used both online and offline Telugu character datasets. To extract the features in the scanned image we used convolutional neural network along with Machine Learning classifiers like Random Forest and Support Vector Machine. Stochastic Gradient Descent (SGD), Root Mean Square Propagation (RMS-P) and Adaptative Moment Estimation (ADAM)optimizers are used in this work to enhance the performance of U-HCR and to reduce the loss function value. This loss value reduction can be possible with optimizers by using CNN. In both online and offline datasets, proposed model showed promising results by maintaining the accuracies with 90.28% for SGD, 96.97% for RMS-P and 93.57% for ADAM respectively.

show abstract

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Unicode-driven Deep Learning Handwritten Telugu-to-English Character Recognition and Translation System

BV Subba Rao

2023

IJRITCC

View full text Add to dashboard Cite

show abstract

“…Programs in Deep Learning usually use more complex capabilities to study, digest, and classify data. DL algorithms have taken the top place in object recognition because they can improve performance [20], [21]. The class of convolutional neural networks (CNN) was first proposed by LeCun [22].…”

Section: Introductionmentioning

confidence: 99%

MikrobatX: The Use of SIFT Feature Extraction in a Deep Learning Approach for Identification and Classification of Microscopic Fragments of Medicinal Leaves

Rahmatulloh,

Suhendy,

Gunawan

2023

Int. J. Adv. Sci. Eng. Inf. Technol.

View full text Add to dashboard Cite

Due to a lack of references that contain standard references, it is still difficult to evaluate the accuracy of the raw material for the medicinal plant Simplicia powder based on microscopic testing in the pharmaceutical industry. Furthermore, it takes much time to manually match the findings of microscopic tests with standard reference materials. For these reasons, artificial intelligence must be used so that researchers can rapidly and reliably forecast the kinds of medicinal plants based on microscopic fragments. Deep learning performance in computer vision has demonstrated encouraging outcomes in recent years. Convolutional neural networks (CNN) enhanced by SIFT feature extraction, dubbed "MikrobatX," are used in the proposed work to identify and classify microscopic fragment images of the medicinal plant Simplicia. This technique plays a key role in the microscopic identification and classification of medicinal plant simplicia. Using microscopic photographs of the leaves of medicinal plants, MikrobatX was able to extract essential Simplicia characteristics. Our proposed model may produce the greatest accuracy value of 89.42% for microscopic medicinal leaf Simplicia image problems, according to experimental results utilizing the Mikrobat dataset. Due to the lack of comparable research using the Microbat dataset, these findings cannot be compared to earlier investigations.

show abstract

“…The online character detection was somewhat easy due to the temporal-based character properties such as procedure, count of distances, stroke, and writing direction. The offline character identification performance was complex due to their difference in fonts and writers [4], [5].…”

Section: Introductionmentioning

confidence: 99%

Puzzle optimization algorithm with capsule networks based Telugu character recognition model

Phaniram

Reddy

2022

IJEECS

View full text Add to dashboard Cite

Recently, <span>Telugu character recognition (TCR) becomes a hot research topic because of drastic increase in technological advancements such as multimedia, and smartphones. Though numerous works have been concentrated on offline TCR models, it is still needed to develop automated and intelligent online TCR models. This paper presents a novel puzzle optimization algorithm with capsule networks-based Telugu character recognition (POACN-TCR) model. The presented POACN-TCR model intends to effectively identify and recognize distinct Telugu characters online. To accomplish this, the POACN-TCR model primarily undergoes pre-processing in different ways such as normalization, smoothing, and interpolation. In addition, the POACN-TCR model designs an effective capsule network (CapsNet) model to generate feature vectors and hyperparameter optimization takes place using puzzle optimization algorithm (POA). Finally, the C4.5 decision tree classifier is utilized for the effectual recognition of Telugu characters. The utilization of POA for hyperparameter optimization of the CapsNet model helps in achieving improved recognition performance. For ensuring the enhanced outcomes of the POACN-TCR model, a wide-ranging experimental analysis is performed and the outcomes pointed out the betterment of the POACN-TCR model on the recent approaches.</span>

show abstract

CNN based Optical Character Recognition and Applications

Cited by 20 publications

References 12 publications

Unicode-driven Deep Learning Handwritten Telugu-to-English Character Recognition and Translation System

Unicode-driven Deep Learning Handwritten Telugu-to-English Character Recognition and Translation System

MikrobatX: The Use of SIFT Feature Extraction in a Deep Learning Approach for Identification and Classification of Microscopic Fragments of Medicinal Leaves

Puzzle optimization algorithm with capsule networks based Telugu character recognition model

Contact Info

Product

Resources

About