eyeSay: Make Eyes Speak for ALS Patients with Deep Transfer Learning-empowered Wearable

Zou, Jiadao; Zhang, Qingxue

doi:10.1109/embc46164.2021.9629874

Cited by 9 publications

(4 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Figure 8 depicts the comparison between the proposed method and other sequential ML methods such as SVM, BiLSTM, and Boosted HMMs using a complex gaze gesture dataset in [59]. Figure 9 depicts the evaluated performance of the proposed methods using Japanese Katakana dataset and other baseline methods such as DTW and HMM [42] and CNN [43]. From Figure 7, our proposed method achieved an average accuracy of 98.81%, precision of 100%, and recall of 98.8%.…”

Section: A Comparison With Some Baseline Methods Using Existing Datasetsmentioning

confidence: 99%

“…Thus, the combination of DNN with HMM outperformed the combination of the Gaussian mixture model (GMM) with HMM in the recognition of Japanese Katakana characters [42]. The DL method improves the recognition performance of eye-writing patterns by using DL layers to automatically extract deep local features [43]. However, DL algorithms require more training datasets to match the performance of ML methods and avoid overfitting during training [14].…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Translated Pattern-Based Eye-Writing Recognition Using Dilated Causal Convolution Network

Bature,

Abdullahi,

Chiracharit

et al. 2024

IEEE Access

View full text Add to dashboard Cite

Recently, eye-writing has been used as a novel language communication method, in which the paths of eye movement are detected for character recognition. However, instability of the eyes causes gaze points to form a character with a non-uniform shape and distinct writing style across participants. The non-uniformity affected the performance of the recognition algorithms and set back the applicability of eyewriting. In this paper root translation and dilated causal convolutional (DCC) layers are utilized to model the non-uniformity in eye-writing patterns. The root translation shifted the pattern to have uniform root gaze points by obtaining the difference between the initial gaze points and subsequent gaze points. The translated patterns were used to train the temporal convolution network (TCN) having three stacked DCC layers with different filter and dilation factors. The DCC layers extract temporal dependencies in the pattern by convolving a particular gaze point with a certain previous gaze point within its receptive field. To evaluate the performance of the proposed method, a dataset of 36 eye-writing characters including 26 English and 10 Arabic numerals was recorded from 20 participants using Tobii eye tracker. The evaluation results depicted that our proposed method achieved an accuracy of 96.20% on our newly designed English and Arabic numeral datasets. The proposed method outperform conventional methods by achieving 98.81%, 97.76%, and 93.51% on HideMyGaze, Complex gaze gesture, and Isolated Japanese Katakana datasets, respectively.

show abstract

Section: A Comparison With Some Baseline Methods Using Existing Datasetsmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Translated Pattern-Based Eye-Writing Recognition Using Dilated Causal Convolution Network

Bature,

Abdullahi,

Chiracharit

et al. 2024

IEEE Access

View full text Add to dashboard Cite

show abstract

“…Machine learning models are not capable of generalizing beyond the circumstances encountered to increase their performance [30]. They were inspired by the human capacity to transfer knowledge, which has led them to focus during training [31], which hampers their ability to on transfer learning to resolve these issues. When compared to the traditional machine learning paradigm, where learning occurs in isolation, not utilizing knowledge from other domains (Figure 5…”

Section: High Level View Of Transfer Learning Paradigmmentioning

confidence: 99%

Comparative Analysis of CNN-Based Smart Pre-Trained Models for Object Detection on DOTA

Hashmi,

Kumar Dwivedi,

Kumar

2024

JAMRIS

View full text Add to dashboard Cite

In this paper, we proposed comparative research on the classification of various objects in satellite images using some pre-trained models of CNN (VGG-16, Inception-V3, ResNet-50, EfficientNet-B7) and R-CNN. In this research work, we have used the DOTA dataset, which combines data from 14 classes. We have implemented above mentioned pre-trained models of CNN, and R-CNN to achieve optimal results for accuracy as well as productivity. To detect objects like ships, tennis courts, swimming pools, vehicles, and harbors from remotely accessed images. In this study, we have used a convolutional neural network (CNN) as the base model. The transfer learning mechanism is employed to speed up the results and for complex computations. We have discovered with the help of experimental analysis that R-CNN and Inception-V3 are performing best out of the five pre-trained models.

show abstract

“…Therefore, a large number of magnetic JPs use Au as micro- and nanomotors to cause changes in motor speed or distance in the presence of targets, generating visual signals [ 122 ]. The asymmetry of JPs can power the movement of particles in complex samples, resulting in visual motion signals in a short period of time [ 207 , 208 ].…”

Section: Janus Particles For Optical Poctmentioning

confidence: 99%

Application of Janus Particles in Point-of-Care Testing

et al. 2022

View full text Add to dashboard Cite

Janus particles (JPs), named after the two-faced Roman god, are asymmetric particles with different chemical properties or polarities. JPs have been widely used in the biomedical field in recent years, including as drug carriers for targeted controlled drug release and as biosensors for biological imaging and biomarker detection, which is crucial in the early detection and treatment of diseases. In this review, we highlight the most recent advancements made with regard to Janus particles in point-of-care testing (POCT). Firstly, we introduce several commonly used methods for preparing Janus particles. Secondly, we present biomarker detection using JPs based on various detection methods to achieve the goal of POCT. Finally, we discuss the challenges and opportunities for developing Janus particles in POCT. This review will facilitate the development of POCT biosensing devices based on the unique properties of Janus particles.

show abstract

eyeSay: Make Eyes Speak for ALS Patients with Deep Transfer Learning-empowered Wearable

Cited by 9 publications

References 13 publications

Translated Pattern-Based Eye-Writing Recognition Using Dilated Causal Convolution Network

Translated Pattern-Based Eye-Writing Recognition Using Dilated Causal Convolution Network

Comparative Analysis of CNN-Based Smart Pre-Trained Models for Object Detection on DOTA

Application of Janus Particles in Point-of-Care Testing

Contact Info

Product

Resources

About