Smartphone-enabled fluorescence sensing platform for As(V) detection in food and environmental matrices using CeO2@Curcumin fluorescent probe.

RCVQA: Visual question answering model based on reading comprehension.

AI-based analysis of label-free live cell imaging of T-cell mediated tumor killing assay enables competitive and robust hit calling.

Segmentation and analysis of OCTA images of coronary heart disease patients based on a multi-dimensional collaborative feature fusion network

Early diagnosis of Alzheimer's disease from functional rs-fMRI images based on deep learning networks and transfer learning approach.

4D flow MRI facilitates quantification of cardiac phase-resolved blood velocity vector fields and has successfully been deployed to study cerebrovascular flow. Besides cardiac-induced flow pulsation, respiration is known to modulate arterial and venous blood flow in the brain. Quantification of the respiratory flow modulation (RFM) holds potential to further our insights into vascular coupling and improve our understanding of cerebral circulation in general. A 5D phase-contrast flow MRI framework was developed to volumetrically quantify RFM by resolving velocity vector fields over the cardiac and respiratory cycle, with high spatial (0.82 mm isotropic) and cardiac (55 ms) resolutions, using two respiratory states whilst accounting for variable physiological RFM delays, with a reasonable acquisition time (20 min at 60 bpm). Recent advances in deep learning-based image reconstruction and analysis methods are incorporated to facilitate the approach. The 5D flow MRI framework was validated in 10 healthy volunteers with reference to fully sampled respiratory-resolved 2D flow MRI orthogonal to the internal carotid artery (ICA), yielding Pearson correlation coefficients of 0.97 and 0.90 and biases of and 0.09 % and 1.77 % for RFM of mean velocity magnitude and amplitude, respectively. The value of cerebrovascular 5D flow MRI is demonstrated using full-field spatially resolved RFM quantification of mean velocity and velocity amplitude, revealing a high physiological intra-subject variability. Cerebrovascular 5D flow MRI enables the study of full-field respiratory flow modulation holding potential of furthering our understanding of cerebral circulation.

VMPicker: A novel cryo-EM particle picker leveraging vision mamba and the segment anything model.

Using the Diffusion Tensor Image Analysis Along the Perivascular Space (ALPS) method and perivascular space (PVS) burden to study glymphatic function in high-risk Alzheimer's disease (AD) populations, this research investigates the correlation between ALPS index and PVS volume with cognitive function respectively. This study enrolled 126 participants, including 21 cognitively unimpaired (CU) individuals, 68 with subjective cognitive decline (SCD), and 37 with mild cognitive impairment (MCI). All participants underwent MRI and cognitive assessments. MRI measures, including the PVS burden and the ALPS index, were compared across SCD, MCI, and CU groups. Additionally, correlations among the ALPS index, PVS burden, and cognitive scales were analyzed. The PVS in the basal ganglia volume fraction (PVSVF-BG) in patients with MCI was significantly larger than the fraction in CUs (p < 0.05) and a higher PVSVF-BG was associated with poorer performance on the Trail Making Test A (TMTA) (r = 0.29, p < 0.05). Compared with the CU and SCD groups, patients with MCI exhibited a significantly lower ALPS index in both the left (p < 0.05) and right hemispheres (p < 0.001). Lower whole brain ALPS index in patients with MCI was correlated with worse performance in the Auditory Verbal Learning Test(AVLT) (N4, r = 0.33, p < 0.05; N7, r = 0.56, p < 0.001). An increased PVS burden and a decreased ALPS index can be observed in the preclinical stage of AD, which may suggest impaired glymphatic system function. These impairments were further correlated with worse cognitive performance in terms of attention in SCD and memory in MCI.

Application of artificial intelligence in colonoscopy imaging for polyp analysis-A systematic review.

Evaluation of seed-metering devices using an automated system based on digital image analysis

The interplay between pore and particle characteristics in sands based on computed tomographic image analyses

Diesel exhaust nanoparticles: Cellular adaptation in lung epithelial and fibroblast cells - An in vitro study.

A comparison of scatter correction software and physical grid in supine chest radiography using a mobile X-ray system across different phantom sizes.

Direct measurement of spreadability and surface quality of powder beds using advanced image analysis for additive manufacturing

Evaluating and enhancing Segment Anything Model transferability for microstructural image analysis in nuclear materials

Subcutaneous tissue structural feature identification using unsupervised machine learning.

Buffalo fly (Haematobia irritans exigua) is recognised for its impact on cattle health, welfare, and production. It is ranked the number one endemic pest for the Australian beef cattle industry by Meat and Livestock Australia. An accurate estimation of fly numbers is essential to evaluate treatment efficacy, phenotyping of susceptible animals for genetic improvement and determining threshold levels to guide integrated pest management strategies. Traditionally, fly numbers are estimated through visual scoring which is inherently challenging as it involves estimating moving flies on a restless host. This study used digital photography and an open-source, semi-automated software package DotDotGoose to count fly numbers on individual animal images. These were then compared to the visual fly scores on the same animals. A random selection of 98 images was used and fly numbers were counted on these images in duplicate by four assessors, one experienced field researcher and three novice assessors. The fly counts on individual images were analysed for consistency and agreement and a consistency and agreement of 99 % was achieved within the four assessors, classified as excellent. The analysis further showed that visual assessments and manual visual counts by experienced assessors consistently underestimated fly numbers compared to the digital image counts. Our results suggest that a digital counting platform offers a more reliable alternative to visual scoring for buffalo fly counts. It improves accuracy and consistency and enables remote image analysis, lowers time and labour costs, and provides the potential for automated real-time monitoring and reporting of fly numbers.

Avalanches present a significant risk in mountainous environments, frequently causing severe trauma and death. While asphyxia is often the primary cause of mortality (65-100 %), the incidence of traumatic injuries, particularly bone fractures, may be increasing due to evolving recreational practices and climate change. This study aims to characterize the distribution and types of fractures in avalanche fatalities. We conducted a retrospective study at Grenoble Alpes University Hospital from April 2023 to April 2025. It included deceased adult avalanche victims with post-mortem computed tomography (CT) scans showing at least one bone fracture. MPR, MIP, and 3D reconstructions were used for image analysis. Thirteen individuals were included (61.5 % male; mean age 37 years). The thorax was the most affected region (92 % of cases). Analysis of 265 fractured bones (79 distinct bones) showed primary fracture distribution in the thorax (52 %), spine (21 %) and skull (14 %). A strong correlation existed between external trauma signs and vital skull lesions, but vital spinal and thoracic lesions often lacked clear external indicators. These preliminary findings contribute to a better understanding of avalanche-specific bone injuries, which could optimize initial management strategies. While extremity fractures are important due to their potential contribution to non-traumatic death and ease of on-site detection, thoracic, spinal and pelvic fractures are challenging to detect in the field. Their accurate identification is yet essential for determining appropriate hospital destination. This exploratory work highlights the need for further larger-cohort studies to establish robust correlations between traumatic mechanisms and avalanche characteristics.

Curvature-aware selective feature interaction network for skin lesion segmentation.

Thermal signatures in breast cancer: Deciphering latent biomarkers through deep learning and explainable AI.