Water pollution classification and detection by hyperspectral imaging

Optics Express, Journal Year: 2024, Volume and Issue: 32(14), P. 23956 - 23956

Published: June 3, 2024

This study utilizes spectral analysis to quantify water pollutants by analyzing the images of biological oxygen demand (BOD). In this study, a total 2545 depicting quality pollution were generated due absence standardized detection method. A novel snap-shot hyperspectral imaging (HSI) conversion algorithm has been developed conduct on traditional RGB images. order demonstrate effectiveness HSI algorithm, two distinct three-dimensional convolution neural networks (3D-CNN) are employed train separate datasets. One dataset is based (HSI-3DCNN), while other (RGB-3DCNN). The categorized into three groups: Good, Normal, and Severe, extent severity. comparison was conducted between models, focusing precision, recall, F1-score, accuracy. model's accuracy improved from 76% 80% when RGB-3DCNN substituted with HSI-3DCNN. results suggest that capacity enhance compared model.

Language: Английский

---

Machine learning–based classification of spatially resolved diffuse reflectance and autofluorescence spectra acquired on human skin for actinic keratoses and skin carcinoma diagnostics aid

Journal of Biomedical Optics, Journal Year: 2025, Volume and Issue: 30(03)

Published: March 4, 2025

The incidence of keratinocyte carcinomas (KCs) is increasing every year, making the task developing new methods for KC early diagnosis utmost medical and economical importance. We aim to evaluate diagnostic aid performance an optical spectroscopy device associated with a machine-learning classification method. present autofluorescence diffuse reflectance spectra obtained in vivo from 131 patients on four histological classes: basal cell carcinoma (BCC), squamous (SCC), actinic keratosis (AK), healthy (H) skin. Classification accuracies by support vector machine, discriminant analysis, multilayer perceptron binary- multi-class modes were compared define best pipeline. accuracy binary tests was >80% discriminate BCC or SCC H. For AK versus other classes, achieved 65% 75% accuracy. In multiclass (three classes) modes, reached 57%. Fusion decisions increased (up 10 percentage point-increase), proving interest multimodal single modality. Such levels are promising as they comparable those general practitioners screening.

Language: Английский

---

Toward near real-time precise supervision of radiofrequency ablation for liver fibrosis using hyperspectral imaging

Spectrochimica Acta Part A Molecular and Biomolecular Spectroscopy, Journal Year: 2025, Volume and Issue: 336, P. 125994 - 125994

Published: March 7, 2025

Language: Английский

---

Leveraging Feature Extraction via T3 Fusion of Deep Learning Models for Enhanced Skin Cancer Classification

Lecture notes in networks and systems, Journal Year: 2025, Volume and Issue: unknown, P. 375 - 385

Published: Jan. 1, 2025

Language: Английский

---

Involution-based HarmonyNet: An efficient hyperspectral imaging model for automatic detection of neonatal health status

Biomedical Signal Processing and Control, Journal Year: 2024, Volume and Issue: 100, P. 106982 - 106982

Published: Oct. 15, 2024

Language: Английский

---

MUCM-Net: a Mamba powered UCM-Net for skin lesion segmentation

Exploration of Medicine, Journal Year: 2024, Volume and Issue: unknown, P. 694 - 708

Published: Oct. 25, 2024

Aim: Skin lesion segmentation is critical for early skin cancer detection. Challenges in automatic from dermoscopic images include variations color, texture, and artifacts of indistinct boundaries. This study aims to develop evaluate MUCM-Net, a lightweight efficient model segmentation, leveraging Mamba state-space models integrated with UCM-Net architecture optimized mobile deployment Methods: MUCM-Net combines Convolutional Neural Networks (CNNs), multi-layer perceptions (MLPs), elements into hybrid feature learning module. Results: The was trained tested on the International Imaging Collaboration (ISIC) 2017 ISIC2018 datasets, consisting 2,000 2,594 images, respectively. Critical metrics evaluation included Dice Similarity Coefficient (DSC), sensitivity (SE), specificity (SP), accuracy (ACC). model’s computational efficiency also assessed by measuring Giga Floating-point Operations Per Second (GFLOPS) number parameters. demonstrated superior performance an average DSC 0.91 ISIC2017 dataset 0.89 dataset, outperforming existing models. It achieved high SE (0.93), SP (0.95), ACC (0.92) low demands (0.055–0.064 GFLOPS). Conclusions: innovative Mamba-UCM layer significantly enhanced while maintaining that suitable devices. establishes new standard balancing exceptional performance. Its ability perform well devices makes it scalable tool detection resource-limited settings. open-source availability supports further research collaboration, promoting advances health diagnostics fight against cancer. source code will be posted https://github.com/chunyuyuan/MUCM-Net.

Language: Английский

---

Machine learning applications in healthcare clinical practice and research

World Journal of Clinical Cases, Journal Year: 2024, Volume and Issue: 13(1)

Published: Nov. 6, 2024

Machine learning (ML) is a type of artificial intelligence that assists computers in the acquisition knowledge through data analysis, thus creating machines can complete tasks otherwise requiring human intelligence. Among its various applications, it has proven groundbreaking healthcare as well, both clinical practice and research. In this editorial, we succinctly introduce ML applications present study, featured latest issue World Journal Clinical Cases . The authors study conducted an analysis using multiple linear regression (MLR) methods to investigate significant factors may impact estimated glomerular filtration rate healthy women with without non-alcoholic fatty liver disease (NAFLD). Their results implicated age most important determining factor groups, followed by lactic dehydrogenase, uric acid, forced expiratory volume one second, albumin. addition, for NAFLD- group, 5th 6th were thyroid-stimulating hormone systolic blood pressure, compared plasma calcium body fat NAFLD+ group. However, study's distinctive contribution lies adoption methodologies, showcasing their superiority over traditional statistical approaches (herein MLR), thereby highlighting potential represent invaluable advanced adjunct tool

Language: Английский

---

The coupling effect between skin strain and blood condition on its reflectance spectrum in-vivo

Optics & Laser Technology, Journal Year: 2024, Volume and Issue: 176, P. 110990 - 110990

Published: April 9, 2024

Language: Английский

---

A proof of concept for microcirculation monitoring using machine learning based hyperspectral imaging in critically ill patients: a monocentric observational study

Critical Care, Journal Year: 2024, Volume and Issue: 28(1)

Published: July 10, 2024

Abstract Background Impaired microcirculation is a cornerstone of sepsis development and leads to reduced tissue oxygenation, influenced by fluid catecholamine administration during treatment. Hyperspectral imaging (HSI) non-invasive bedside technology for visualizing physicochemical characteristics. Machine learning (ML) skin HSI might offer an automated approach assessment, providing individualized fingerprint critically ill patients in intensive care. The study aimed determine if machine could be utilized automatically identify regions interest (ROIs) the hand, thereby distinguishing between healthy individuals with using HSI. Methods raw data from 75 30 controls were recorded TIVITA® Tissue System analyzed ML approach. Additionally, divided into two groups based on their SOFA scores further subanalysis: less severely (SOFA ≤ 5) > 5). analysis was fully-automated MediaPipe ROI detection (palm fingertips) feature extraction. Features statistically highlight relevant wavelength combinations Mann–Whitney-U test Benjamini, Krieger, Yekutieli (BKY) correction. In addition, Random Forest models trained bootstrapping, importances determined gain insights regarding importance model decision. Results An pipeline generating ROIs extraction successfully established. accurately distinguished patients. Wavelengths at fingertips differed ranges 575–695 nm 840–1000 nm. For palm, significant differences observed range 925–1000 Feature plots indicated information same ranges. Combining palm fingertip provided highest reliability, AUC 0.92 distinguish controls. Conclusion Based this proof concept, integration standardized along analyzes, able differentiate sepsis. This offers reliable objective assessment microcirculation, facilitating rapid identification

Language: Английский

---

HGLeNet DEEP LEARNING FOR SKIN CANCER DETECTION WITH SKIN LESION SEGMENTATION

Journal of Mechanics in Medicine and Biology, Journal Year: 2024, Volume and Issue: unknown

Published: July 18, 2024

Skin cancer is deemed to be the most dangerous type of cancer. It occurs due damage caused DNA and if left untreated can lead death. Various methods have been devised over last few years for skin detection. However, their performance was affected by various challenges in image analysis, like color illumination variations, differences shape, size, etc. Therefore, tackle these issues novel framework, deep learning (DL) technique accurate detection lesion segmentation developed. Primarily, pre-processed employing a weighted median filtering eradicate noises contained. Then, lesions carried out efficient neural network (ENet). After that, augmentation accomplished avoid overfitting, later, feature extraction out. At last, effectuated with help hybrid GoogleNet-LeNet (HGLeNet), which obtained merging GoogleNet LeNet, here layers are modified using regression concept. Furthermore, introduced framework examined its effectiveness through accuracy, sensitivity specificity. Moreover, HGLeNet attained highest accuracy 0.922, 0.928, as well specificity 0.924.

Language: Английский

---