Meta loses bid to end BIPA voiceprint suit as case moves forward A federal judge in California rejected Meta Platforms’ attempt to end a biometric privacy lawsuit over voice recordings collected through Facebook and Messenger, finding that evidence about Meta’s own voice processing technologies and its ability to connect audio files with user accounts creates a factual dispute over whether the recordings could be used to identify people. Illinois resident Natalie Delgado, who brought the lawsuit in 2023 under the state’s Biometric Information Privacy Act (BIPA), filed a motion August 21 asking U.S. District Judge Susan Illston of the Northern District of California to certify an Illinois class. A hearing on class certification is scheduled for December 18. BIPA expressly protects “voiceprints” as biometric identifiers, along with fingerprints, retina or iris scans and scans of hand or face geometry. But the statute does not define what makes recorded speech a voiceprint rather than simply an audio recording. Meta argued that distinction should end Delgado’s case. According to the company, the audio files it received through Facebook and Messenger were merely voice recordings. To identify a speaker, Meta argued, a recording would first have to be processed into different data representing characteristics of the speaker’s voice, which it called an “Output Representation.” That representation would then have to be compared with data linking it to a known person. Meta said it had not created such representations from Delgado’s recordings and did not possess information linking the voices in her uploaded recordings to her identity. Illston concluded the evidence was not that simple. The central question under BIPA, she said, was not whether Meta had used Delgado’s voice recordings to identify her. Citing a 2024 Ninth Circuit ruling involving Meta’s facial recognition technology, she said the relevant issue is whether the biometric
Aug 25, 2026 · via biometricupdate.com
Amnesty: Argentina’s AI Surveillance is Social Control - Jurist News - 6 hours ago - 3 min read August 21, 2026 Amnesty International on Wednesday accused Argentina of building techno-authoritarian surveillance infrastructure without adequate legal safeguards, as the government’s newly created AI security unit expands state monitoring powers over citizens’ online communications and physical movements. In its latest report, the rights group stated that Argentina‘s Ministry of Security spent at least USD$1.2 million on surveillance technology between 2024 and 2025, which included social media monitoring tools, facial recognition software, and thermal-camera drones with automatic tracking and ability to transmit images in real time. Amnesty expressed concern for the surveillance of protesters, journalists, activists, human rights defenders, migrants, young people, and marginalized groups. They warn that this can create a “chilling effect,” in which people who believe they are being monitored may avoid speaking, protesting or organizing, even without being directly arrested or punished. This raises concerns about the impact of surveillance on the right to freedom of expression, freedom of peaceful assembly, and freedom of association. In October 2025, Argentina bought a Clearview AI facial recognition license for USD$33,500. Clearview’s database contains more than 70 billion facial images scraped from the internet, and data protection authorities in France, Italy, Greece, and the Netherlands have already fined the company over its data practices. The Ministry, however, maintains that the Federal Police use facial recognition only in connection with active judicial investigations, and no evidence has surfaced that Clearview was used specifically to identify protesters. Argentina’s expansion of AI-driven surveillance has largely occurred through executive and ministerial measures. On July 26, 2024, then-Security Minister Patricia Bullrich signed Resolution 710/2024, creating the Artificial Intelligence Unit Applied to Security (UIAAS) inside the Ministry of Security’s Directorate of Cybercrime and Cyber Affairs. The resolution authorizes
Aug 25, 2026 · via genocidewatch.com
Key takeaways from this investigation - County prosecutors in Illinois fed sensitive personal and location information about local residents — including court dates and scheduled jail releases — to federal immigration agents. - The practice was widespread: Prosecutors’ offices in 1 out of every 6 counties in Illinois provided information to ICE. - Illinois’ TRUST Act prohibits local law enforcement agencies from providing information to federal immigration officials, but state’s attorneys disagreed about whether their offices are bound by the law. - The consequences were devastating for families: Some people who didn’t appear to otherwise be on ICE’s radar were deported after county prosecutors shared information about them with federal agents. This story was produced in partnership with WIRED. Rolando Perez Samayoa had come to the United States in 2023 seeking safety. Fleeing violence in Mexico, he’d first planned to make a life in Alabama but had been told that Illinois would protect immigrants like him. “There’s some kind of law there,” he said. “It’s supposed to be a sanctuary.” He was referring to the state’s 2017 TRUST Act, which was meant to keep local law enforcement out of federal deportation efforts. Samayoa made his way to southern Illinois and found work cleaning a factory. After a DUI arrest in Marion County last October, Samayoa said, he attended his hearings — a fact confirmed by court records. In late January, John Christeson, an assistant state’s attorney in the office prosecuting the case, emailed an agent with Homeland Security Investigations, a division of U.S. Immigration and Customs Enforcement, about Samayoa. The message included his date of birth, an upcoming court date, and where he lived: Centralia, Illinois. “Ok, we’ll take care of it, thank you,” the federal agent, Sergio Fulgencio, wrote back minutes later. Two days later, he wanted to know
Aug 25, 2026 · via injusticewatch.org
Mass Surveillance Safety Not Surveillance Manchester Townhall Join the #SafetyNotSurveillance coalition in Manchester as we explore how AI and ‘crime-predicting’ technologies are currently being used by Greater Manchester Police, the harms and risks of their continues roll out, and what action local communities and councils can take to invest in true, lasting community safety – not surveillance. - WHEN Thursday 24 September, 6-9pm BST - WHERE Central Hall (Collier Room), Oldham Street, Manchester M1 1JQ - SPEAKERS Sarah Lasoye (Open Rights Group) Dr Patrick Williams Kids of Colour Northern Police Monitoring Project Manchester Townhall on AI and ‘crime-predicting’ technologies In the last year, we’ve seen dramatic escalations in the use of AI tools across UK policing. In January, the Home Office announced its plans to invest £115 million into PoliceAI – accelerating the roll out of AI technologies across all 43 forces in England and Wales. These range from administrative tools, to flawed live facial recognition cameras and so called ‘crime-prediction’ and ‘risk-scoring’ systems that rely on algorithmic models. Just this month, we’ve seen GMP deploy facial recognition at Manchester Caribbean Carnival. We’ve seen police forces push ahead with these technologies, justifying their use with the paternalistic argument that the presumed benefits of embedding surveillance across public life, outweigh any consequent harms to the public. What we don’t want, is policing without fully-informed consent by the public, and particularly, racialised and marginalised communities whose views and voices are routinely sidelined. Refreshments and food will be provided 🥘 Manchester Townhall Register Now If you have questions, or would like support in order to fully access and participate this event, please email Sarah, our Pre-Crime Programme Manager.
Aug 25, 2026 · via openrightsgroup.org
Police used live facial recognition (LFR) technology during a recent operation at The Lexicon.
On Tuesday, August 18, Thames Valley Police deployed LFR vans to identify known suspects, deter criminal activity, and enhance public safety.
Thames Valley Police said: "The specialist LFR team will be working alongside our local officers to identify known suspects, deter crime and help keep our communities safer for everyone.
"Preventing crime is just as important as catching those who break the law."
Officers notified the public about the deployment in advance to maintain transparency.
Thames Valley Police said: "While this can mean some individuals choose to avoid us, the vans can help deter crime and keep the area safer."
Members of the public were encouraged to speak with officers to learn about the technology.
Share
Aug 25, 2026 · via bracknellnews.co.uk
Council weighs new rules for Flock camera data amid citywide surveillance concerns HUNTINGTON, W.Va (WCHS) — Implementing guardrails on the use of Flock cameras was on the agenda on Monday night at Huntington city council. This comes at a time where people citywide have voiced concerns over the systems. noting instances where the data could be easily accessed by anyone. "It is nothing more than mass surveillance. The data stored they can say seven days. I don’t believe a word of that,” Huntington resident, Trish Lilly said. Council members heard the first reading of an ordinance that would place rules and guidelines on how the data could be gathered from the flock cameras. Many people who live in Huntington continue to voice their opposition, making it clear they do not want flock in their community. "We're not doing facial recognition. This technology with flock cannot do facial recognition it does not track you as a person it takes a picture of the back of your car it's not used for traffic stops for rolling through red lights rolling through a stop sign. It’s not for traffic enforcement; that's against state law. It's not watching where you go to church can't see who your doctor is whether you go into the ER. It's not it's pointed a public street at an intersection,” Mayor Patrick Farrell said during Monday’s meeting. Councilmember Rumbaugh highlighted the importance of adding criminalization into the ordinance on the table. "What harm would there be to have the municipal criminalization back in place? Specifically, I would love to re-include and amend to have the language that you had before in your policy back in so that we have a little bit of teeth, a light adjustment a light penalty,” Tia Rumbaugh said during Monday’s meeting. The ordinance will
Aug 25, 2026 · via wchstv.com
Another Pyramid Scheme word search is live! Play today's puzzle here!
Need a little head start? Today's hint: WINE
Sign up here to get notified whenever a new puzzle goes live. 📬
HOW TO PLAY: Find every hidden word by connecting neighboring letters. Words can be built in any direction — even diagonally — and every letter is used exactly once. There's only one winning combination. 🧠
A new, challenging word search arrives every day. Come back tomorrow for another puzzle!
Can't wait? Get caught up by exploring the archive!
Comments
Aug 25, 2026 · via buzzfeed.com
Storms that once seemed exceptional are showing up more often, and the added-up destruction they cause is growing. A new study suggests one practical way to limit some of that damage may already be in place through tougher building standards that make homes less vulnerable to hurricane winds. Here's what to know Researchers from Kyoto University and Kobe University described in a press release how they created a large-scale way to detect wind damage from hurricanes by scanning drone and satellite imagery taken after storms for blue tarps. Those tarps are commonly used to cover damaged roofs. Using images collected after Hurricane Irma struck Florida in September 2017, they estimated both where wind damage occurred and how severe it was. That damage information was then linked to building-permit data for more than 1 million homes across eight counties in South Florida. To assess the impact of Florida's 2001 building code, the study compared homes constructed just before the new rules took effect with those built just after. The code required stronger roof-to-wall connections, thicker roof sheathing, and, in some areas, impact-resistant windows and doors. In the analysis, homes built under the newer code were about 22% less likely to sustain wind damage, and the damage they did experience was roughly 27% less severe. "This study provides large-scale evidence that stronger building codes are effective for reducing wind risk and adapting to climate change," commented co-author Kenji Takeuchi, in the release. More background When hurricanes damage homes, the losses generally come from flooding and wind. After Hurricane Ian in 2022, 44% of insurance claims were tied to flooding and 56% to wind damage. Even with that split, researchers have often devoted more attention to flooding because elevation data and hydrology models make flood risk easier to estimate. Tracking wind damage over
Aug 25, 2026 · via thecooldown.com
Who is watching Mid-Michigan? Inside the use of Flock license plate readers An investigation into how Lansing and Jackson police use automatic license plate readers, and the growing debate over oversight LANSING, Mich. — Automatic license plate readers have been surveilling Mid-Michigan roads for years, and a review of how the technology is used by the region’s two largest city police departments shows differences in policy, oversight and public trust. The cameras, made by Flock Safety, are mounted on utility poles, street lamps and traffic lights. They are subtle enough that most drivers would not give them a second glance, but they scan nearly every car that passes by. Flock cameras are in use in 50 Michigan counties. Customers include private businesses and law enforcement agencies, including the Lansing and Jackson police departments. Both departments say their goal is to solve crime faster. How the technology works The camera captures an image of a passing car. That image uploads to a searchable database. Officers can then search by plate, vehicle type, or an identifier like a cracked windshield. Lansing Police Chief Rob Backus said his officers use Flock data to find missing people, track down traffickers and catch thieves. “We’re always looking for what that next component of our investigations is going to be. So as we saw license plate readers start to be more advanced and more applicable to investigations, we definitely looked into that,” Backus said. “This is almost like a needle in a haystack,” Backus told News 10 Investigates. “We need to connect it to a person. We need to connect it to a location. That’s how we’re going to make progress in solving these crimes.” Jackson Police Chief Chris Simpson said the technology saves time during investigations. “Instead of relying or trying to comb through hours
Aug 24, 2026 · via wilx.com
NYC authorities have yet to nab ‘mole people’ in manhole next to Penn Station last week: ‘Reconnaissance, 9/11’ See more of our coverage in your search results. Add The New York Post on Google The NYPD on Monday said no busts have yet been made involving the “mole people’’ who climbed from a manhole just feet from the main entrance to Penn Station last week — terrifying the public. “I just immediately thought, reconnaissance, 9/11,” said an Arizona tourist named Mark to The Post in packed Herald Square in Manhattan. “If you’re an Iranian soldier, you need to cripple New York City? Under Penn Station? Under Madison Square Garden? Aren’t all the train lines crisscrossing here?” A trio of underground scofflaws was caught on video around 4:20 a.m. Aug. 18 wearing headlamps and overalls and pulling themselves out of a sanitation and drainage manhole about 150 feet from Penn Station — just a block from Madison Square Garden — then hurriedly walking away. A worker with nearby Penn 2 developer Vornado Realty Trust said he was aware that cops had been snooping around the tourist- and business-choked area collecting potential surveillance footage of the creepy nightcrawlers. “The more they told me about it, the more I wanted to get transferred to another area,” the man said. The disturbing incident was eerily similar to two “mole people” sightings in Brooklyn earlier in the year, which led to a handful of arrests involving people who cops believe were hunting for valuables. Police have collected surveillance footage that possibly captured the latest incident to try to ID the three people using facial recognition technology and track them down to figure out what they were up to, law-enforcement sources said. Cops only became aware of the incident Sunday after other video of the “mole
Aug 24, 2026 · via nypost.com
GEO donations coincide with expanding ICE detention and biometric monitoring contracts A subsidiary of GEO Group, whose ICE business spans detention, biometric monitoring and location intelligence, gave more than $1.4 million in July to a political committee supporting President Donald Trump. The payments by GEO Reentry Services came within days of GEO announcing two new ICE detention contracts expected to generate a combined $165 million in their first full year of operations. A Biometric Update review of ICE procurement records found that one of those agreements is substantially larger than GEO’s public description suggested. The government estimates the five-year requirement for GEO’s Rivers Facility in North Carolina at nearly $457.1 million and awarded it without full and open competition after receiving seven responses to earlier market research. Another wholly owned GEO subsidiary, BI Incorporated, operates ICE’s nationwide Intensive Supervision Appearance Program (ISAP) using electronic monitoring, biometric reporting and case management to supervise people on the agency’s non-detained docket. BI also holds one of ICE’s new skip-tracing contracts, allowing it to use government case information and commercial data to help locate people ICE is trying to find. The political spending was first reported by The Lever and subsequently by The New York Times after MAGA Inc. disclosed its July receipts to the Federal Election Commission (FEC). GEO Reentry gave MAGA Inc. $1 million on July 16, three days after GEO announced a five-year agreement for the 1,188-bed Big Horn Facility in Hudson, Colorado. It gave another $413,000 on July 31, two days after GEO announced its agreement for the 1,320-bed Rivers Facility in Winton, North Carolina. The July payments were not GEO Reentry’s first to MAGA Inc. FEC records and campaign finance disclosures show the subsidiary gave the committee $1 million in October 2025, another $1 million in March and another
Aug 24, 2026 · via biometricupdate.com
BRENHAM — The Brenham Police Department has been awarded a state grant to fund key public safety initiatives, including the deployment of additional Automated License Plate Reader (ALPR) cameras and a first-responder drone platform. The grant, totaling approximately $81,666 from the State of Texas Motor Vehicle Crime Prevention Authority for the fiscal year 2027 Catalytic Converter Theft Prevention Grant Program, was formally accepted following a unanimous vote by the Brenham City Council. The department will contribute a matching requirement of $16,334, funded through its asset forfeiture account. Police Chief Gary Boshears presented the grant details during a recent council meeting, highlighting that the application ranked near the top of over 100 statewide submissions. The funding will support the installation of six to seven new ALPR units placed on mobile platforms and strategic roadways. These specialized cameras capture high-speed images of license plates to help law enforcement locate stolen vehicles, track suspects in active investigations, and detect vehicles tied to violent crimes or catalytic converter thefts. During the meeting, Councilmember Steve Soman inquired about data privacy, user management, and the potential integration of artificial intelligence or third-party databases with the new systems. Chief Boshears clarified that the department is working with technology provider Axon and emphasized that the system does not utilize facial recognition technology. "The LPRs do not use facial recognition," he explained. "They can build a small profile on the vehicle itself. I.E. it's a red Ford F-150, but it doesn't use any type of facial recognition. It's specifically designed to take a picture of the license plate on the rear of the vehicle." Addressing concerns regarding national databases and unauthorized searches, Chief Boshears emphasized that data sharing is strictly controlled at the local level. "By default, Axon's data sharing is turned off for everyone, and so we can
Aug 24, 2026 · via southtexasnews.com
A Munich court has dealt a significant legal setback to Suno, ruling on Friday that the AI music company violated German copyright law by training its model on songs without securing licenses from the rights holders. The decision, issued by the Munich Regional Court, hands a major victory to GEMA,... Read More » Twitch Streamer Sues Amazon Over AI Training, Claims Creator Content Used Without Consent A Twitch streamer has sued Twitch and Amazon in a proposed class action, claiming the companies used creators’ videos, chats and other content to train generative artificial intelligence models without their consent or compensation. Warren Pandiscia, a Connecticut creator who has streamed on Twitch for nearly a decade, filed the lawsuit Aug. 20 in the U.S. District Court for the Northern District of California. He claims Twitch gave Amazon access to creator content for AI training even though the platform’s agreements with users did not authorize that use. The lawsuit followed an Aug. 12 announcement from Twitch introducing a setting that allows creators to opt out of having their channel content used to train Amazon’s generative AI models. The setting is enabled by default, requiring creators to turn it off if they do not want their content used for future AI training. Pandiscia argues the alleged use began before Twitch disclosed the new setting. His complaint claims the companies had already used creator material for AI work without first obtaining permission and benefited from content produced by streamers at their own expense. Twitch Chief Product Officer Mike Minton acknowledged in 2024 that Amazon was using Twitch content in some AI work, which he described at the time as prototyping rather than production use. During an Aug. 12 Twitch discussion, Minton said he did not know what Amazon had previously used for model training. Pandiscia
Aug 24, 2026 · via lawcommentary.com
The Santa Barbara City Council on Tuesday will discuss the termination of its contract with Flock Safety amid national backlash and concerns over privacy vulnerabilities. Though it won’t be the end of automatic license plate readers in Santa Barbara. While the city’s contract with Flock Safety has ended, the city will propose a replacement to the cameras from a different company: Verkada. Verkada, a California based security company, is the expected new provider of the automatic license plate readers. The City Council is set to vote on Tuesday at its 2 p.m. meeting. City Councilmember Meagan Harmon told the News-Press that she has a “deep discomfort” with surveillance cameras such as Flock Safety ones. “However, cameras in general have been a hugely important tool in our ability to open and close criminal cases and quite literally solve crimes,” Harmon said. “I’m very glad that the community conversation is happening, there are real costs and consequences with video surveillance but there are also huge benefits.” She said that the new proposal involving Verkada is an improvement from the Flock Safety cameras. “Opposed to the Flock contract, which was externally arranged, all of our footage now will be held within our own system. To me that is quite a meaningful difference,” Harmon said. What is Verkada? With headquarters in San Mateo, California, and regional hubs in nine other U.S. cities, Verkada offers a variety of security services. The company also has international offices in eight countries including Australia, Taiwan, and Poland. Verkada’s camera technology uses artificial intelligence to aid its security systems. AI is used to “locate footage of suspects or vehicles in seconds with powerful AI analytics,” wrote the company in their product description for “Smart solutions for city and county security”. The company recently announced Hard Rock Stadium, home of
Aug 24, 2026 · via newspress.com
You may have seen the widespread backlash over the recent popularity of smartglasses, such as those from Meta, which feature tiny cameras that can take still photos and videos of people without their knowledge, and project a heads-up display of information onto the lenses visible only to the wearer. They have already been banned by many theatres, restaurants and pubs and have now been banned from courts in England and Wales too. Sleazy owners have been quick to exploit their potential for creating sexualised content of women in the street, who they approach, chat up and secretly film. This has in some cases led to unsuspecting victims being named in the comments and even their addresses shared. In one case a user tried to blackmail the victim when she asked him to remove a video of her that he had posted on social media. No wonder they have already been dubbed ‘pervert glasses’, and the pressure group Everyone Hates Elon has been putting up billboard ads around London showing them being worn by Jeffrey Epstein with the tagline ‘Glasses for people who don’t do consent’. From the point of view of Meta, these smartglasses are essentially mass data collection and AI training devices. Which is fine for surveillance-state fans who want to share every aspect of their lives with Mark Zuckerberg and his tech billionaire buddies (including going to the toilet and other intimate moments, according to the poor Kenyans paid little over a dollar an hour to watch all this footage for Meta) but the people being unknowingly filmed by these glasses as they go about their day have not consented to being part of the program. Some defenders of the new technology point to how useful they are for blind and partially sighted people and claim that the
Aug 24, 2026 · via amateurphotographer.com
The data center political attacks are here As public sentiment on data centers continues to sour, both Democratic and Republican candidates are taking a more oppositional stance to them. • 3 min read TL;DR: There might be nothing more politically unpopular right now than being a data center cheerleader. Politicians on both sides of the aisle are taking note—and changing their stance—as the midterm elections approach, while AI companies are desperately trying to buy back some goodwill. What happened: In an interview this past Sunday, Texas Gov. Greg Abbott declared that AI companies “dug their own grave” amid the data center backlash they’re facing—weeks after he ordered a moratorium on new data center power approvals in the state. It’s a big shift in tone from a governor who called Texas the “epicenter of AI development” last November. And he’s far from alone: An analysis from The Washington Post shows that American politicians have gotten a lot more negative about data centers since last year. Republican leaders have typically been more welcoming of them than their Democratic counterparts, but a few pro-data center elected officials have either lost their elections or faced recalls in recent months. Now, politicians from both parties are adjusting their stance and ramping up attack ads accusing opponents of being too data center-friendly. AI’s expensive charm offensive: AI firms, meanwhile, are probably feeling the walls closing in on them. In a private memo that leaked last week, the National Republican Senatorial Committee told AI companies to fix their PR problem, saying they had a “toxic brand.” Tech Brew breaks down the biggest tech news, emerging innovations, workplace tools, and cultural trends so you can understand what's new and why it matters. By subscribing, you accept our Terms & Privacy Policy. To be fair, the industry has been
Aug 24, 2026 · via techbrew.com
Which are the best cameras, lenses and photographic accessories you can buy right now? The answer can be found right here, in the EISA Awards 2026-2027, as voted for by the magazines and websites that form the photography group within the Expert Imaging and Sound Association (EISA). EISA is a group of 51 special-interest magazines and websites from 26 countries around the world, covering Hi-Fi, photography, mobile devices, in-car electronics, and home theatre and vision. Each member is a renowned expert within its subject area and known for the rigour and integrity of its testing. As one of the organisation’s founding members, Amateur Photographer is an integral part of EISA’s Photography Panel. We work closely alongside 15 other magazines from Europe, India and the USA, all of which are among the most respected in their home countries. To ensure that the winning products are truly worthy of your hard-earned money, each one must have been tested in depth by multiple panel members. We don’t give awards based on quick first looks, or on the basis of specifications alone. While each country has its own particular preferences for which features and characteristics are the most important in different kinds of cameras and lenses, there is often a strong consensus as to which should be recognised with an award. Read on to discover the winners of the EISA Awards 2026-2027 – the best photography products of the year! EISA Camera of the Year 2026-2027 Sony Alpha 7 V The Sony Alpha 7 V is a weather-sealed full-frame camera that’s suitable for virtually any application. It combines tried-and-tested features such as 5-axis image stabilisation with cutting edge photographic technology. Its partially stacked Exmor RS CMOS image sensor delivers 33 megapixels while enabling shooting at 30 frames per second to provide blackout free continuous
Aug 24, 2026 · via amateurphotographer.com
Label-free imaging offers considerable benefits compared to traditional fluorescence-based methods, including faster processing, reduced reagent costs, and the ability to monitor live cells. Label-free detection also offers key advantages by allowing observation of biomolecular interactions without sample modification, resulting in more accurate data, deeper characterisation, and real-time monitoring of cellular dynamics. This article outlines the use of customisable artificial intelligence (AI) models within IN Carta® image analysis software to develop deep-learning-based segmentation models suitable for accurate cell and nuclear detection that only use transmitted-light (TL) images. Label-free analysis was initially established using the U2OS cell line. A pre-trained AI model was first applied to segment nuclei in Hoechst-stained images, enabling the development of a TL-based nuclear segmentation model. The resulting segmentation masks were paired with corresponding TL images and used to train a new model able to segment nuclei using only TL images. This approach enabled label-free nuclear detection with over 97% accuracy, as validated against Hoechst-based nuclear counts. The model was also able to reliably quantify both damaged and live cell phenotypes. The model was further trained using images from additional cell lines, including HCT116, U2OS, HeLa, MCF7, and HEK. The model achieved more than 90% detection accuracy across the majority of these cell types. The five cell lines were treated with 10 anti-cancer compounds using a seven-point dilution series to evaluate the model’s utility in compound screening. Cell counts acquired via TL-based nuclear segmentation were compared against conventional nuclear staining methods across treatments. The results showed strong correlation and comparable IC50 values, verifying the label-free approach’s accuracy and reliability. This label-free approach offers important advantages in streamlined processing speed and workflow simplicity, allowing real-time analysis of compound effects. This makes the approach a powerful tool in high-throughput and phenotypic screening applications. Label-free detection and analysis offer a range
Aug 24, 2026 · via news-medical.net
Abstract Brain tumour classification from Magnetic Resonance Imaging (MRI) requires both high diagnostic accuracy and computationally efficient model adaptation. Although deep transfer learning has improved automated tumour recognition, most existing approaches remain limited by dependence on generic ImageNet-pretrained weights, prolonged fine-tuning time, and weak institution-specific adaptability under small local medical datasets. To address these limitations, this paper proposes an Adaptive Incremental Domain Pretraining and Frozen-Weight Local Rapid Adaptation framework built on the ResNet18 backbone. Unlike conventional one-step transfer learning, the proposed method recursively preserves and updates tumour-specialised parameter states across multiple same-domain MRI repositories, progressively constructing a domain-adapted diagnostic backbone. The final inherited model is then subjected to frozen-weight local rapid adaptation for institution-specific MRI customisation. Experimental evaluation shows that direct ImageNet-based ResNet18 transfer learning achieves 93.27. Similar content being viewed by others Subjects Introduction Brain tumours constitute one of the most critical neurological disorders because of their direct influence on the central nervous system and their potentially life-threatening progression. Uncontrolled abnormal growth of intracranial or spinal cells can severely impair physiological regulation, cognitive activity, and neural communication, thereby resulting in substantial morbidity and mortality. Clinically, brain tumours are broadly categorised as benign or malignant, with benign tumours generally slow-growing and localised, whereas malignant tumours are highly invasive, aggressive, and rapidly progressive1. According to the World Health Organization (WHO) classification, low-grade tumours (Grades I and II) exhibit comparatively better prognosis, whereas high-grade tumours (Grades III and IV) are associated with extremely poor survival outcomes, thereby making early-stage identification and precise grading an indispensable requirement for effective treatment planning and prognosis estimation2,3. Among the various diagnostic modalities available for tumour identification, including computed tomography (CT), electroencephalography (EEG), and invasive biopsy-assisted examinations, Magnetic Resonance Imaging (MRI) has emerged as the most reliable and widely preferred non-invasive clinical tool for brain tumour assessment
Aug 24, 2026 · via nature.com
Market Data and Widgets Technology provided by TradingView
Loading symbol info...
Market Data and Widgets Technology provided by TradingView. Data Delayed by 10 minutes. Information is provided 'as is' and solely for informational purposes, not for trading purposes or advice.
Aug 24, 2026 · via kitco.com