Predicting lifestyle using BioVRSea multi-biometric paradigms DOI
Marco Recenti, Deborah Jacob, Romain Aubonnet

et al.

2022 IEEE International Conference on Metrology for Extended Reality, Artificial Intelligence and Neural Engineering (MetroXRAINE), Journal Year: 2022, Volume and Issue: 46, P. 329 - 334

Published: Oct. 26, 2022

BioVRSea was recently introduced as an unique multi-biometric system that combine Virtual Reality with a moving platform to induce Motion Sickness (MS). Electromyography (EMG) and balance features measuring the center of pressure (CoP) are among bio-signals measured during six segments protocol on BioVRSea. A total 262 participants has been all them underwent MS questionnaire self-assess relative symptoms personal information like smoking, physical activity Body Mass Index. From last three data binary lifestyle index is created Machine Learning models used classify it starting from EMG CoP groups taken individually together. After appropriate feature's selection, multiple algorithms applied best results for classification reached K Nearest Neighbors algorithm (0.83 maximum accuracy 0.60 recall) while Random Forest perform AUCROC (0.64). The most relevant ones second segment experiment, before movements, its first light movements. These show unhealthy influences in negative way performance person term induced task. They can also be preliminary input study influence behavior people who suffers serious problems or neuro-degenerative patients using novel platform.

Language: Английский

Effect of X-ray scatter correction on the estimation of attenuation coefficient in mammography: a simulation study DOI
Mario Sansone, Alfonso Maria Ponsiglione, Francesca Angelone

et al.

2022 IEEE International Conference on Metrology for Extended Reality, Artificial Intelligence and Neural Engineering (MetroXRAINE), Journal Year: 2022, Volume and Issue: 4, P. 323 - 328

Published: Oct. 26, 2022

A mammographic image requires high contrast for soft tissue imaging. Even small amounts of dispersion reduce the required to make accurate diagnoses. Current systems digital mammography use an anti-scatter grid scatter phenomenon. However, despite widespread grids in clinical practice, it leads elimination useful primary radiation, thus forcing increase patient irradiation order achieve images. It is therefore desirable develop processing methods correction. The objective this study evaluate how effective removal can be achieved by implementing and tuning appropriate deconvolution functions means a simulation approach carried out on rectangular breast phantoms, with ultimate aim proposing framework evaluation comparison between experimental theoretical attenuation coefficient as indirect measure scattering effects mammography. phantom composed two types step blocks representing adipose glandular breast, provided manufacturer. In study, assumed that measured result convolution (devoid scattering) spatially variant Point Spread Function, which represents scattered radiation. allows recovery assessment impact phenomenon coefficients examined sample.

Language: Английский

Citations

4
Machine Learning and Biosignals are able to discriminate biomechanical risk classes according to the Revised NIOSH Lifting Equation DOI
Leandro Donisi, Giuseppe Cesarelli, E Capodaglio

et al.

2022 IEEE International Conference on Metrology for Extended Reality, Artificial Intelligence and Neural Engineering (MetroXRAINE), Journal Year: 2022, Volume and Issue: unknown

Published: Oct. 26, 2022

Many work activities can imply a biomechanical overload. Among these activities, lifting loads may determine work-related musculoskeletal disorders. To limit injuries, the National Institute for Occupational Safety and Health (NIOSH) proposed methodology to assess risk in tasks through an equation based on intensity, duration, frequency other geometrical characteristics of tasks. In this work, we explored feasibility tree-based machine learning algorithms classify according Revised NIOSH equation). Electromyography signals acquired from biceps sternum acceleration collected during were registered using wearable sensor (BITalino (r)evolution) worn by 5 healthy young subjects. segmented as extract region interest related actions and, each interest, several time domain features extracted. Interesting results obtained terms evaluation metrics binary risk/no-risk classification. conclusion, indicates combination represents valid approach automatically equation. Future investigation enriched study populations could confirm capabilities potential risky activities.

Language: Английский

Citations

4
Evaluating the Feasibility of AI-Predicted mpMRI Image Features for Predicting Prostate Cancer Aggressiveness: a Multicenter Study DOI
Kexin Wang,

Ning Luo,

Zhaonan Sun

et al.

Research Square (Research Square), Journal Year: 2024, Volume and Issue: unknown

Published: April 25, 2024

Abstract Objective To evaluate the feasibility of utilizing artificial intelligence (AI)-predicted multiparametric MRI (mpMRI) image features for predicting aggressiveness prostate cancer (PCa). Materials and methods A total 878 PCa patients from 4 hospitals were retrospectively collected, all whom had pathological results after radical prostatectomy(RP).A pre-trained AI algorithm was used to select suspected lesions extract lesion model development. The study evaluated five prediction methods, including 1) clinical selected by algorithm, 2)the PIRADS category, 3)a conventional radiomics model, 4) a based on deep learning, 5)biopsy pathology. Results In externally validated dataset, learn-based showed highest area under curve (AUC 0.700 0.791).It exceeded 0.597 0.718), traditional radiomic 0.566 0.632), score 0.554 0.613) biopsy pathology 0.537 0.578). And AUC predicted did not show statistically significant difference among three verified (P > 0.05). Conclusion Deep-radiomics models AI-extracted mpMRI images can potentially be predict aggressiveness, demonstrating generalized ability external validation.

Language: Английский

Citations

0
Comparative analysis of SVM and k-nearest neighbor classification algorithm in fingerprint detection DOI Creative Commons

R. Sravanthi,

A. Gnana Soundari

AIP conference proceedings, Journal Year: 2024, Volume and Issue: 3097, P. 020129 - 020129

Published: Jan. 1, 2024

Views Icon Article contents Figures & tables Video Audio Supplementary Data Peer Review Share Twitter Facebook Reddit LinkedIn Tools Reprints and Permissions Cite Search Site Citation R. Sravanthi, A. Gnana Soundari; Comparative analysis of SVM k-nearest neighbor classification algorithm in fingerprint detection. AIP Conf. Proc. 7 May 2024; 2853 (1): 020129. https://doi.org/10.1063/5.0197527 Download citation file: Ris (Zotero) Reference Manager EasyBib Bookends Mendeley Papers EndNote RefWorks BibTex toolbar search Dropdown Menu input auto suggest filter your All ContentAIP Publishing PortfolioAIP Conference Proceedings Advanced |Citation

Language: Английский

Citations

0
Predicting lifestyle using BioVRSea multi-biometric paradigms DOI
Marco Recenti, Deborah Jacob, Romain Aubonnet

et al.

2022 IEEE International Conference on Metrology for Extended Reality, Artificial Intelligence and Neural Engineering (MetroXRAINE), Journal Year: 2022, Volume and Issue: 46, P. 329 - 334

Published: Oct. 26, 2022

BioVRSea was recently introduced as an unique multi-biometric system that combine Virtual Reality with a moving platform to induce Motion Sickness (MS). Electromyography (EMG) and balance features measuring the center of pressure (CoP) are among bio-signals measured during six segments protocol on BioVRSea. A total 262 participants has been all them underwent MS questionnaire self-assess relative symptoms personal information like smoking, physical activity Body Mass Index. From last three data binary lifestyle index is created Machine Learning models used classify it starting from EMG CoP groups taken individually together. After appropriate feature's selection, multiple algorithms applied best results for classification reached K Nearest Neighbors algorithm (0.83 maximum accuracy 0.60 recall) while Random Forest perform AUCROC (0.64). The most relevant ones second segment experiment, before movements, its first light movements. These show unhealthy influences in negative way performance person term induced task. They can also be preliminary input study influence behavior people who suffers serious problems or neuro-degenerative patients using novel platform.

Language: Английский

Citations

2