Articles tagged with Image Analysis
Researchers from the University of Groningen and Spain developed a method to train AI systems using distractions to improve image recognition. By analyzing how deep learning systems process images, they found that forcing the system's focus towards secondary characteristics can lead to better performance.
A new study reveals that AI tools analyzing tumor images can make inaccurate predictions due to shortcuts introduced by institutions submitting the images. The models often rely on location-specific signatures rather than biological characteristics, leading to biased outcomes for patients from certain medical centers.
Researchers developed a non-contact method using thermal images to detect delayed healing of venous leg ulcers. The study shows textural analysis can predict whether a wound needs extra management as early as week two.
A team of researchers led by Zili Shen and Pieter van Dokkum used Hubble Space Telescope to measure the distance of ultra-diffuse galaxy NGC1052-DF2, confirming it lacks dark matter. The results are based on 40 orbits of the telescope and provide crucial implications for estimating the physical properties of the galaxy.
The Chan Zuckerberg Initiative (CZI) has awarded nearly $28 million in grants to support the development of next-generation electron microscopy techniques for visualizing proteins in cells. The awards will enable researchers to obtain unprecedented views of protein structure, quantity, distribution, and interactions at near-atomic reso...
A team of researchers at the University of Texas at Austin has applied a machine learning algorithm to analyze Synthetic Aperture Radar images and create detailed maps of land types. This new approach can help predict storm surge risk and inform mitigation strategies, such as building 'green walls' to protect inland areas.
A new artificial intelligence tool can analyze patients' stool images to help diagnose chronic gastrointestinal issues such as IBD and IBS. The Smart Toilet technology provides a more accurate and timely diagnosis by collecting long-term data on stool form and presence of blood.
Researchers have developed a cloud-based platform called ZeroCostDL4Mic that uses artificial intelligence to analyze microscopy images, improving image quality and detecting features such as tumors in biopsy samples. The platform is freely available online and requires no coding expertise, making it accessible to non-experts.
Smartphone-based imaging systems can guide diagnosis and treatment with portable, user-friendly biomedical imaging. Emerging technologies like multispectral and quantitative fluorescence imaging offer promising diagnostic applications.
A study at the University of Helsinki uses machine learning to analyze bone marrow samples and identify genetic mutations associated with myelodysplastic syndrome (MDS). The technique provides new insights into disease progression and can help assess and diagnose hematological malignancies.
A new toolkit developed by researchers at the University of Houston, in collaboration with the National Institutes of Health, enables rapid drug development for brain diseases and injuries. The toolkit uses high-resolution imaging to analyze unhealthy brain tissue, revealing multiple biological processes at once.
A large-scale study found that the introduction of RADLogics' AI-Powered solution into radiology workflow reduced report turnaround time by an average of 30 percent, equivalent to 7 minutes per case. The study also showed a significant improvement in productivity and report accuracy.
Researchers introduce a novel algorithm to estimate speed and angle of rotating objects in space, then apply these estimates to develop high-resolution images. The method achieves improved resolution without being heavily affected by atmospheric fluctuations.
GraphIt, a state-of-the-art graph programming language, has been extended to run on graphics processing units (GPUs), enabling faster graph analysis. This advance accelerates graph algorithms, particularly those that benefit from parallelism, such as recommendation algorithms and internet search functions.
Researchers have uncovered evidence of what may be the earliest-known use of symbols, dating back to around 120,000 years ago. The discovery was made on a bone fragment in central Israel and features six similar etchings that are believed to hold symbolic or spiritual significance.
Researchers found that aging leads to increased intrinsic ignition and decreased metastability in the whole-brain functional network. This suggests a deficiency in efficient global communication in the brain, contributing to cognitive decline.
A scanning transmission X-ray microscope for lithium analysis has been developed, enabling the study of chemical states and behaviors with high spatial resolution. The microscope uses a low-pass filtering zone plate to suppress higher-order harmonics, allowing for accurate measurement of XAS spectra.
Researchers from Kumamoto University have developed a highly sensitive method to evaluate cytoskeleton organization from microscopic images. The new technique uses computer simulations and image analysis to accurately measure bundle states, even in unclear images.
A study investigated the correlation between ADC parameters, Ki-67 expression, and clinical outcomes in PCNSL patients. In immunocompetent patients, ADC histogram analysis predicted tumor proliferation and survival. However, in HIV-positive patients, ADC values showed limited utility in predicting clinical outcomes.
The journal features papers on AI in medicine, including new roles for clinicians and AI-aided disease prediction. It also covers the development of chest imaging analysis methods for COVID-19 diagnosis.
Researchers pooled data from multiple sites to visualize neuroinflammation in ALS patients, enhancing detection and developing drugs. The study used two second-generation translator protein tracers and found improved power to detect inflammation compared to individual site data alone.
The article reviews AI-assisted chest imaging analysis methods for COVID-19 diagnosis, demonstrating the potential of deep learning-based techniques. Establishing a database for public research and extracting lesions accurately are crucial to improve performance.
A University of Georgia researcher used computer vision to analyze thousands of images from over 100 Instagram accounts of United States politicians, discovering that posts featuring politicians' faces in non-political settings attract more likes and comments. The study found that images with only the politician's face or in personal s...
A novel method for automated pollen analysis has been developed by combining imaging flow cytometry with deep learning, allowing for accurate species identification and quantitative findings in just 20 minutes. The new tool was tested on 35 plant species and achieved an accuracy rate of 96%, outperforming traditional microscopy methods.
Experts have documented and analyzed hitherto unknown details in the muqarnas of the Lions' Courtyard at the Alhambra, revealing a different configuration and number of pieces between the two temples. The research also detected geometric deformations due to centuries-old repairs.
Researchers are combining different types of brain imaging data to capture patterns indicative of brain disorders. The goal is to develop a compact framework to map brain disorders, providing a signature that tells us how mental illnesses impact the brain.
Chemical imaging modalities have helped reveal the Eyckian Lamb of God's hidden facial features, including v-shaped nostrils and forward-gazing eyes. A smaller, more naturalistically shaped body with a sagging back and rounded hindquarters has also been discovered, suggesting revisions to the original design.
Researchers at Ruhr-University Bochum developed a new method for efficiently identifying deep-fake images by analyzing objects in the frequency domain. The study found that images generated by GANs exhibit artefacts in the high-frequency range, which can be used to distinguish them from real photos.
Researchers have developed a new method to analyze microscopic samples without using external light, reducing interference and damage to living specimens. The 'glow in the dark' approach uses chemical stimuli to activate chemicals, enabling precise control over localized oxidative hotspots.
A team has devised deep-learning and other computational approaches that dramatically reduce image-analysis time by orders of magnitude. They report their results in Nature Biotechnology, accelerating image analysis in three major ways: deconvolution, 3D registration, and complex deconvolution.
Rice University scientists developed an easy and affordable tool to count and characterize nanoparticles using the open-source SEMseg program. The tool simplifies nanoparticle analysis by extracting pixel-level data from low-contrast SEM images and recombining it into sharp images.
KAIST researchers developed an algorithm for automated 3D brain imaging data analysis, enabling precise and quantitative mapping of complex neural circuits. The new technology allows for accurate comparison of brain data from different animals and improves the accuracy and consistency of analysis results.
Researchers have developed a new imaging analysis pipeline called LaMDA, which enables machine learning and complex analyses on high-dimensional microscopy data. This can reveal and predict T-cellular states, providing valuable insights for drug testing, vaccine development, and expanding knowledge of T-cell biology.
The U.S. Army Research Laboratory has developed a novel parallel computing framework, GPOP, to process large graphs and data efficiently, making it ideal for DOD security analysis and applications such as terrorist-tracking and bioinformatics.
A team of scientists has developed a system utilizing image analysis and artificial intelligence to analyze the shape of large numbers of seeds from a single image. The trained model detected and segmented individual seeds with high accuracy and analyzed seeds of other crops, accelerating crop breeding and analysis.
A Brazilian research group uses VideometerLab, a multispectral imaging instrument, to analyze pupae and detect dead or empty insects. The device provides precise data on sample quality, improving the efficiency of biological control.
Researchers Juliette Shedd and Sheherazade R. Jafari are developing a graduate-level conflict resolution training course for Burma Leadership Activity scholars. The five-year project aims to improve technical competencies and leadership among participants, promoting respect for ethnic and religious diversity.
A new AI algorithm developed by researchers can analyze retinal images to predict which patients with diabetic macular edema are likely to respond to anti-VEGF therapy. The algorithm achieved an 87% accuracy rate in predicting treatment response, offering a potential breakthrough in personalized medicine.
A study analyzing 190 fMRI studies found that food and erotic images activate the left hemisphere, while money activates both hemispheres. The basal ganglia play a key role in reward processing, with different regions activated depending on the type of reward.
Susan Allen received $10,000 from GMU Foundation for her research at the Center for Peacemaking Practice. The funding supports her work on conflict resolution and peacemaking practice.
Tropical Storm 06A is maintaining its strength in the Arabian Sea, with powerful storms indicating heavy rainfall potential. The storm is expected to make landfall in east central Somalia late on Dec. 6.
Researchers reestimated China's inland waterbodies using satellite images from 1984-2015, finding a 0.2-1.5 times increase in surface area and 0.3-1.5 times higher numbers of larger waterbodies than previous estimates.
Scientists from Skoltech ADASE lab have developed a method to enhance depth map resolution, enabling more realistic virtual reality and computer graphics. The new approach leverages human perception to assess reconstruction quality, outperforming existing methods in visual quality.
Claudio Silva was inducted into the inaugural cohort of the IEEE Visualization Academy for his seminal contributions to point-based modeling and surface reconstruction. He is a highly respected researcher with numerous publications and patents, including those related to VisTrails and urban data visualization.
A KAIST research team developed an advanced system to automatically analyze visual information in films and quantify gender bias. The study found that female characters are often depicted in passive emotions and surrounded by specific objects, promoting public awareness of the importance of prudence when creating characters.
Masashi Watanabe, a Lehigh University professor, is recognized by the Microanalysis Society for his research on materials characterization using various electron microscopy approaches. He has developed innovative techniques such as the zeta-factor method and multivariate statistical analysis.
QUT researchers have developed a new method for analysing Optical Coherence Tomography (OCT) images using deep learning techniques to detect and track eye diseases such as glaucoma and macular degeneration. The team's findings show that their program is reliable and more accurate than standard image analysis methods.
NASA's Terra satellite revealed the strongest thunderstorms in Tropical Storm 14W were located east of the center, fueled by extremely cold cloud top temperatures. The storm is moving westward and is expected to intensify to 90 knots over the next five days.
The study analyzed 135 million imaging examinations among 16-21 million patients in the US and Canada, finding increased medical imaging rates among adults but declining nuclear medicine use. Imaging rates continued to rise among children except for CT, which declined since 2006.
Tropical Storm 14W strengthened into a powerful system due to strong thunderstorms fueled by infrared data from NASA's Aqua satellite. The storm's intense cloud top temperatures indicate potential for heavy rainfall, with winds near 40 mph and expected to intensify over the next five days.
NASA's Terra satellite detected scattered cold clouds in Hurricane Henriette, indicating a weakening system. The storm's cloud top temperatures reached as low as -50 degrees Fahrenheit, further supporting its decline.
Researchers developed a simplified new mass spectrometric technique using a continuous wave laser and graphene substrate to analyze bio samples without sample preparation. This technology can obtain high-resolution images and secure enough heat needed for specimen analysis with small amount of light generated by the continuous wave laser.
A new standard, BIDS, allows researchers to compare and combine studies in an apples-to-apples way. This enables the sharing and analysis of large brain imaging datasets, improving the field's ability to understand brain function.
Tropical Storm Francisco is expected to bring heavy rainfall to southern Japan as it approaches landfall. The strongest storms have cloud top temperatures as cold as minus 70 degrees Fahrenheit, indicating strong potential for heavy rainfall.
Tropical Storm Wipha formed quickly in the South China Sea, with NASA's Aqua satellite detecting strong thunderstorms and cold cloud top temperatures. The storm is expected to move northwest towards southern China, making landfall near Hanoi, Vietnam.
NASA's Terra satellite reveals two areas of intense storms in Tropical Storm Nari, with cloud top temperatures as low as minus 50 degrees Fahrenheit. The storm is moving northward toward Japan, expected to make landfall and then dissipate over cooler waters.
NASA's Aqua satellite detected a small area of strong thunderstorms with cold cloud top temperatures, enough to maintain Tropical Depression Dalila's tropical cyclone classification. The depression is forecasted to weaken and become a post-tropical remnant low in the coming days.
The switch from 2D to 3D cell cultures in drug discovery improves the accuracy of human physiological conditions and increases the success rates of drugs. High-content analysis with 3D cell cultures is expected to increase the quality of compounds progressing to preclinical stages.
A new system uses a single labeled scan and unlabeled scans to automatically generate a large dataset of distinct training examples. The system's approach learns anatomical, brightness, and contrast variations from unlabeled scans to synthesize realistic and accurately labeled scans.
The discovery of rare marine fossil specimens in northern Portugal has provided a more detailed picture of biodiversity during the Middle Ordovician period. Soft-bodied specimens have filled a gap in understanding evolution during this time, suggesting that many soft-bodied fossils remain to be discovered.