Azmain Yakin Srizon

Mr. Azmain Yakin Srizon

Rajshahi University of Engineering & Technology
orcid

Research Focus & Metrics

Academic Career Length

8Years (2019-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
31 August 2021 in Sensors
Verified
Journal Article
Open access
PDF file
Jungpil SHINJungpil SHIN
Jungpil SHIN
AMAkitaka Matsuoka
Akitaka Matsuoka
Hasan Md. Al MehediHasan Md. Al Mehedi
Hasan Md. Al Mehedi
Azmain Yakin SrizonAzmain Yakin Srizon
Azmain Yakin Srizon
Sign language is designed to assist the deaf and hard of hearing community to convey messages and connect with society. Sign language recognition has been an important domain of research for a long time. Previously, sensor-based approaches have obtained higher accuracy than vision-based approaches. Due to the cost-effectiveness of vision-based approaches, researchers have been conducted here also despite the accuracy drop. The purpose of this research is to recognize American sign characters using hand images obtained from a web camera. In this work, the media-pipe hands algorithm was used for estimating hand joints from RGB images of hands obtained from a web camera and two types of features were generated from the estimated coordinates of the joints obtained for classification: one is the distances between the joint points and the other one is the angles between vectors and 3D axes. The classifiers utilized to classify the characters were support vector machine (SVM) and light gradient boosting machine (GBM). Three character datasets were used for recognition: the ASL Alphabet dataset, the Massey dataset, and the finger spelling A dataset. The results obtained were 99.39% for the Massey dataset, 87.60% for the ASL Alphabet dataset, and 98.45% for Finger Spelling A dataset. The proposed design for automatic American sign language recognition is cost-effective, computationally inexpensive, does not require any special sensors or devices, and has outperformed previous studies.

29 April 2022 in IEEE Access
Verified
Journal Article
Open access
KHKoki Hirooka
Koki Hirooka
Hasan Md. Al MehediHasan Md. Al Mehedi
Hasan Md. Al Mehedi
Jungpil SHINJungpil SHIN
Jungpil SHIN
Azmain Yakin SrizonAzmain Yakin Srizon
Azmain Yakin Srizon
Human activity recognition is one of the most difficult tasks in computer vision. Due to the lack of time information, detecting human activities from still photos is more difficult than sensor-based or video-based techniques. Recently, various deep learning based solutions are being proposed one after another, and their performance is constantly improving. In this paper, we proposed a convolutional neural architecture by ensembling transfer learning based multi-channel attention networks. Here, four CNN branches were used to make feature fusion based ensembling and in each branch, an attention module was used to extract the contextual information from the feature map produced by existing pre-trained models. Finally, the extracted feature maps from four branches were concatenated and fed to fully connected network to produce the final recognition output. We considered 3 different datasets, Stanford 40 actions, BU-101 and Willow human actions datasets to evaluate our system. Experimental analysis showed that the proposed ensembled convolutional architecture outperformed previous works by a noteworthy margin.

25 July 2021 in Applied Sciences
Verified
Journal Article
Open access
PDF file
ASAbu Sayeed
Abu Sayeed
Jungpil SHINJungpil SHIN
Jungpil SHIN
Hasan Md. Al MehediHasan Md. Al Mehedi
Hasan Md. Al Mehedi
Azmain Yakin SrizonAzmain Yakin Srizon
Azmain Yakin Srizon
MHMd Mehedi Hasan
Md Mehedi Hasan
As it is the seventh most-spoken language and fifth most-spoken native language in the world, the domain of Bengali handwritten character recognition has fascinated researchers for decades. Although other popular languages i.e., English, Chinese, Hindi, Spanish, etc. have received many contributions in the area of handwritten character recognition, Bengali has not received many noteworthy contributions in this domain because of the complex curvatures and similar writing fashions of Bengali characters. Previously, studies were conducted by using different approaches based on traditional learning, and deep learning. In this research, we proposed a low-cost novel convolutional neural network architecture for the recognition of Bengali characters with only 2.24 to 2.43 million parameters based on the number of output classes. We considered 8 different formations of CMATERdb datasets based on previous studies for the training phase. With experimental analysis, we showed that our proposed system outperformed previous works by a noteworthy margin for all 8 datasets. Moreover, we tested our trained models on other available Bengali characters datasets such as Ekush, BanglaLekha, and NumtaDB datasets. Our proposed architecture achieved 96–99% overall accuracies for these datasets as well. We believe our contributions will be beneficial for developing an automated high-performance recognition tool for Bengali handwritten characters.

28 November 2020
Verified
Conference Paper
Closed access
MHMehedi Hasan
Mehedi Hasan
Azmain Yakin SrizonAzmain Yakin Srizon
Azmain Yakin Srizon
ASAbu Sayeed
Abu Sayeed
AHAl Mehedi Hasan
Al Mehedi Hasan
Having a massive community of almost 466 million deaf-mute people all over the world, sign language recognition has always fascinated researchers to develop sophisticated models that can successfully recognize sign languages. Because of not being a universal language, sign language differs in terms of languages and communities. Previously, various researches have been conducted on different sign languages. In this study, we considered the Sign Language MINST dataset. Previously, different classifiers like support vector machine, random forest, multilayer perceptron, etc. have been introduced for sign language recognition. Recently, shallow CNN and Capsule Networks have obtained better results. Therefore, in this research, we proposed a deep convolutional neural network model to achieve the successful identification of the sign linguistics alphabets. After implementing the model, we produced an overall accuracy of 97.62% and comparison with previous researches revealed that our proposed model outperformed all previously introduced models.

1 January 2020
Verified
Conference Paper
Closed access
MHMehedi Hasan
Mehedi Hasan
Azmain Yakin SrizonAzmain Yakin Srizon
Azmain Yakin Srizon
AHAl Mehedi Hasan
Al Mehedi Hasan
With approximately 466 million deaf-mute people in the world, sign language recognition has been an area of interest for researches for a long time now. Previously, many researchers have conducted studies on different sign language datasets, achieving a decent accuracy of recognition. However, not many studies have been conducted on the Bengali sign alphabet recognition dataset despite having approximately 3 million severe deaf cases alone. Previous studies on Bengali sign character detection suggested decent accuracies using traditional machine learning approaches. However, in recent studies, some researches have achieved higher test accuracy using deep learning models. In our study, we've proposed a novel Convolutional Neural Network (CNN) model for the recognition of the Bengali sign alphabets from the Ishara-Lipi dataset. Our model achieved an overall accuracy of 99.22% that has outperformed all previous studies of Bengali sign character recognition for the deaf-mute community.

6 April 2021
Verified
Conference Paper
Closed access
MHMehedi Hasan
Mehedi Hasan
Azmain Yakin SrizonAzmain Yakin Srizon
Azmain Yakin Srizon
ASAbu Sayeed
Abu Sayeed
AHAl Mehedi Hasan
Al Mehedi Hasan
Having the community with more than 500,000 deaf and mute English speaking people, sign alphabet detection has become a domain of interest among the researchers for the last decade. Previously, a lot of signs of progress have been made in accurate recognition of the American Sign Language. Both convolutional neural networks and traditional machine learning classifiers have been applied formerly for the recognition process. In this study, we've considered an American Sign Language dataset having 36 classes of English characters and digits. The latest research on this real-world dataset achieved an accuracy of 90%. However, in our research, we introduced modified inceptionV3 architecture for the detection of American Sign characters and obtained overall correctness of 98.81% which outperformed all the previous studies by a notable margin.