A real-time hand-signs segmentation and classification system using fuzzy rule based RGB model and grid-pattern analysis

Rahaman, Muhammad Aminur; Jasim, Mahmood; Ali, Md. Yousuf; Zhang, Tao; Hasanuzzaman, Md.

doi:10.1007/s11704-018-7082-4

Cited by 10 publications

(5 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Though they claim to have an accuracy of 98.99% in detecting the BdSL, their approach is not applicable for real-time recognition. M. A. Rahaman et al in [11], present a real-time Bengali and Chinese numeral signs recognition system using contour matching. The system is trained and tested using a total 2000 contour templates separately for both Bengali and Chinese numeral signs from 10 signers and achieved recognition accuracy of 95.80% and 95.90% with a computational cost of 8.023 milliseconds per frame.…”

Section: Related Workmentioning

confidence: 99%

“…The available BdSL datasets are not sufficient enough [11,12,17,18,19] to develop a fully functioned real-time BdSL detection system. Half of these datasets [11,12,17,20] are also not available for further research. These datasets are further built on a controlled lab environment and the only exception is [18] but this dataset only contains 10 classes.…”

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

“…But, to train a deep learning classifier, the role of a large dataset is unavoidable and plays a pivotal role to build a robust classifier. However, deep learning methods, so far, on BdSL letters recognition are restricted to constraint or small scale dataset [11,12,13], which is not useful to build a robust classifier. The collection of sign letter data is challenging due to the lack of open-source information available on the internet, human resources, the deaf community and knowledge.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

BdSL36: A Dataset for Bangladeshi Sign Letters Recognition

Hoque¹,

Jubair²,

Akash³

et al. 2021

Computer Vision – ACCV 2020 Workshops

View full text Add to dashboard Cite

Bangladeshi Sign Language (BdSL) is a commonly used medium of communication for the hearing-impaired people in Bangladesh. A real-time BdSL interpreter with no controlled lab environment has a broad social impact and an interesting avenue of research as well. Also, it is a challenging task due to the variation in different subjects (age, gender, color, etc.), complex features, and similarities of signs and clustered backgrounds. However, the existing dataset for BdSL classification task is mainly built in a lab friendly setup which limits the application of powerful deep learning technology. In this paper, we introduce a dataset named BdSL36 which incorporates background augmentation to make the dataset versatile and contains over four million images belonging to 36 categories. Besides, we annotate about 40, 000 images with bounding boxes to utilize the potentiality of object detection algorithms. Furthermore, several intensive experiments are performed to establish the baseline performance of our BdSL36. Moreover, we employ beta testing of our classifiers at the user level to justify the possibilities of real-world application with this dataset. We believe our BdSL36 will expedite future research on practical sign letter classification. We make the datasets and all the pre-trained models available for further researcher.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

BdSL36: A Dataset for Bangladeshi Sign Letters Recognition

Hoque¹,

Jubair²,

Akash³

et al. 2021

Computer Vision – ACCV 2020 Workshops

View full text Add to dashboard Cite

show abstract

“…However, their recognition was not in real time. In [11], the authors also introduced an approach for detecting BdSL letters and digits which applies a fuzzy-logic based model and gridpattern analysis in real-time. In [12], the authors presented a real-time Bengali and Chinese numeral signs recognition system using contour matching.…”

Section: A Existing Bdsl Detection Techniquesmentioning

confidence: 99%

Real Time Bangladeshi Sign Language Detection using Faster R-CNN

Hoque

Jubair

Islam

et al. 2018

2018 International Conference on Innovation in Engineering and Technology (ICIET)

View full text Add to dashboard Cite

Bangladeshi Sign Language (BdSL) is a commonly used medium of communication for the hearing-impaired people in Bangladesh. Developing a real time system to detect these signs from images is a great challenge. In this paper, we present a technique to detect BdSL from images that performs in real time. Our method uses Convolutional Neural Network based object detection technique to detect the presence of signs in the image region and to recognize its class. For this purpose, we adopted Faster Region-based Convolutional Network approach and developed a dataset -BdSLImset -to train our system. Previous research works in detecting BdSL generally depend on external devices while most of the other vision-based techniques do not perform efficiently in real time. Our approach, however, is free from such limitations and the experimental results demonstrate that the proposed method successfully identifies and recognizes Bangladeshi signs in real time.

show abstract