Improvement of Classification Accuracy of Four-Class Voluntary-Imagery fNIRS Signals using Convolutional Neural Networks

Milu, Md. Mahmudul Haque; Rahman, Md. Asadur; Rashid, M. A.; Kuwana, Anna; Kobayashi, Haruo

doi:10.48084/etasr.5703

Cited by 11 publications

(5 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As a result, it is referred to as a False forged. This situation highlights the value of accurate categorization and identification of natural features while also highlighting the limitations of predictive algorithms [16][17][18].…”

Section: B Results On Testing Datamentioning

confidence: 99%

Digital Image Forensics: An Improved DenseNet Architecture for Forged Image Detection

Alzahrani

2024

Eng. Technol. Appl. Sci. Res.

View full text Add to dashboard Cite

Images sent across internet platforms are frequently subject to modifications, including simple alterations, such as compression, scaling, and filtering, which can mask possible changes. These modifications significantly limit the usefulness of digital image forensics analysis methods. As a result, precise classification of authentic and forged images becomes critical. In this study, a system for augmented image forgery detection is provided. Previous research on identifying counterfeit images revealed unexpected outcomes when using conventional feature encoding techniques and machine learning classifiers. Deep neural networks have been also utilized in these efforts, however, the gradient vanishing problem was ignored. A DenseNet model was created to tackle limitations inherent in typical Convolutional Neural Networks (CNNs), such as gradient vanishing and unnecessary layer requirements. The proposed DenseNet model architecture, which is composed of densely connected layers, is designed for precise discrimination between genuine and altered images. A dataset of forged images was implemented to compare the proposed DenseNet model to state-of-the-art deep learning methods, and the results showed that it outperformed them. The recommended enhanced DenseNet model has the ability to detect modified images with an astonishing accuracy of 92.32%.

show abstract

Section: B Results On Testing Datamentioning

confidence: 99%

Digital Image Forensics: An Improved DenseNet Architecture for Forged Image Detection

Alzahrani

2024

Eng. Technol. Appl. Sci. Res.

View full text Add to dashboard Cite

show abstract

“…Huma Hamid et al [38] presented a study to compare the ML classifiers (SVM, K-NN, and LDA) with DL (CNN, LSTM, and Bi-LSTM) algorithms to perform the classification of two classes of walk and rest tasks and reported better performance for DL algorithms. A similar study presented by Mahmudul Haque Milu et al [39] applied ML (support vector machine (SVM) and linear discriminant analysis (LDA)) and compared it with CNN; they reported that CNN performed well in automatic feature extraction as compared to ML. The conventional LSTM algorithm classifies signals based on the input, forget, and output gate mechanism.…”

Section: Discussionmentioning

confidence: 96%

Enhancing Classification Accuracy with Integrated Contextual Gate Network: Deep Learning Approach for Functional Near-Infrared Spectroscopy Brain–Computer Interface Application

Akhter,

Naseer,

Nazeer

et al. 2024

Sensors

View full text Add to dashboard Cite

Brain–computer interface (BCI) systems include signal acquisition, preprocessing, feature extraction, classification, and an application phase. In fNIRS-BCI systems, deep learning (DL) algorithms play a crucial role in enhancing accuracy. Unlike traditional machine learning (ML) classifiers, DL algorithms eliminate the need for manual feature extraction. DL neural networks automatically extract hidden patterns/features within a dataset to classify the data. In this study, a hand-gripping (closing and opening) two-class motor activity dataset from twenty healthy participants is acquired, and an integrated contextual gate network (ICGN) algorithm (proposed) is applied to that dataset to enhance the classification accuracy. The proposed algorithm extracts the features from the filtered data and generates the patterns based on the information from the previous cells within the network. Accordingly, classification is performed based on the similar generated patterns within the dataset. The accuracy of the proposed algorithm is compared with the long short-term memory (LSTM) and bidirectional long short-term memory (Bi-LSTM). The proposed ICGN algorithm yielded a classification accuracy of 91.23 ± 1.60%, which is significantly (p < 0.025) higher than the 84.89 ± 3.91 and 88.82 ± 1.96 achieved by LSTM and Bi-LSTM, respectively. An open access, three-class (right- and left-hand finger tapping and dominant foot tapping) dataset of 30 subjects is used to validate the proposed algorithm. The results show that ICGN can be efficiently used for the classification of two- and three-class problems in fNIRS-based BCI applications.

show abstract

“…The findings of the object detection phase, along with associated information, are subsequently forwarded to the final phase of the model. Within this phase, an integrated proximity detection algorithm [27][28] calculates the distance between recognized objects and the visually impaired user. This model also records nearby objects, along with essential details such as the date and time of identification, during the transition from textual information to voice-based output in the text-to-voice conversion phase [25].…”

Section: Methodology Computer Vision Is An Advanced Technology That O...mentioning

confidence: 99%

A Binary Object Detection Pattern Model to Assist the Visually Impaired in detecting Normal and Camouflaged Faces

Sajini,

Pushpa

2024

Eng. Technol. Appl. Sci. Res.

View full text Add to dashboard Cite

This study presents a novel Binary Object Detection Pattern Model (BODPM) to detect objects with face key points and recognize them using the KERAS dataset. The proximity and accuracy of the recognized items were evaluated using computer vision techniques. The object recognition time interval and duration were recorded and stored permanently in a database, and the information was communicated to the visually impaired user as voice output. The normal face, without wearing a mask, was identified using binary patterns with proximity detection. Camouflaged objects were detected in a maximum probability range of 100%. The proposed method was tested, calculating accuracy and score, and compared with existing models, showcasing remarkable performance. The proposed method of normal and camouflage detection is a novel prediction with proximity analysis of objects in a frame.

show abstract

Improvement of Classification Accuracy of Four-Class Voluntary-Imagery fNIRS Signals using Convolutional Neural Networks

Cited by 11 publications

References 22 publications

Digital Image Forensics: An Improved DenseNet Architecture for Forged Image Detection

Digital Image Forensics: An Improved DenseNet Architecture for Forged Image Detection

Enhancing Classification Accuracy with Integrated Contextual Gate Network: Deep Learning Approach for Functional Near-Infrared Spectroscopy Brain–Computer Interface Application

A Binary Object Detection Pattern Model to Assist the Visually Impaired in detecting Normal and Camouflaged Faces

Contact Info

Product

Resources

About