Last Updated on 07/01/2026 by Chris Gampat Sony photographers have been spoiled for choices. The company has introduced some fantastic cameras over the years, and it seems that they are now focusing on smaller sensor devices. After the launch of the R1XR Mk III, a luxury compact camera for professionals, it seems that Sony has something in store for birding and wildlife photography enthusiasts. Have a look. After almost a decade of silence, Sony is said to introduce the RX10 V. Sony Alpha Rumor (SAR) reported that the camera is set to launch in July. The two things that the portal confirmed, in addition to the date, are a new, larger NP-FZ100 battery and a new image processing chip. The battery has been used in Sony’s Alpha mirrorless lineup, and a new battery could help with one of the most common complaints: shorter battery life. In fact, a new battery can also help with the support of BIONZ XR chips, which can help with better AI-driven subject recognition autofocus and real-time tracking. It can also fulfill video needs, which is ideal for some hybrid users. While SAR has not revealed any specifications, one can also assume that the camera will house the same 20.1-megapixel stacked 1-inch Exmor RS sensor. The changes will largely be how the images are processed. In addition, the iconic 24-600mm lens from the 9-year-old RX10 IV is also expected to remain the same. Until some concrete evidence is introduced, it is best to assume that the camera can have limited changes. For instance, the A1 II featured the same sensor, but the AI chip helped to get a better autofocus for photographers who photograph people. While we have not tested the RX10 IV, we did make a brief first impression of the camera. The Sony RX10
Jul 1, 2026 · via thephoblographer.com
Elon Musk, briefly the world’s first trillionaire – but now a mere billionaire again – is a man of exceptions. He’s built not one, but two of the world’s most pioneering technology companies (Tesla and SpaceX). He was talking about settling humans on Mars with a straight face some 20 years ago. Unlike most tech CEOs, he posts on social media multiple times daily, via his own platform, X. In 2025, he gave what looked like a Nazi salute, very publicly, in Washington DC. That same year, he held a very senior role in the United States government, with no prior political experience, while simultaneously expanding his business empire. In his brief and chaotic tenure as head of the Department of Government Efficiency (DOGE), he tried to turn government into a problem of data synthesis and pattern recognition, leading to optimised policy solutions. All the while, he seemed to forget that real people, entitled to fairness and justice, were affected profoundly by his desk-based decisions. All this has made him a household name and one of the world’s most powerful individuals. Some, like journalist Cory Doctorow, have been asking: is he now exceptionally dangerous? And where does he fit in with other oft-criticised West Coast “broligarchs”, like Amazon’s Jeff Bezos, Palantir’s Alexander Karp and Meta’s Mark Zuckerberg? Review: Muskism: A Guide for the Perplexed – Quinn Slobodian and Ben Tarnoff (Allen Lane) To answer these questions, you need to scrutinise both the man and the means at his disposal. This is exactly what Canadian political economist Quinn Slobodian and technology journalist Ben Tarnoff do in their carefully researched, well written and thought-provoking book, Muskism: A Guide for the Perplexed. “Muskism” is a reference to “Fordism”, named after industrialist and motor vehicle manufacturer Henry Ford, whose mass production model altered American
Jun 30, 2026 · via theconversation.com
ICE biometrics underpin broader surveillance network, report argues A coalition of privacy and civil liberties organizations says that biometric systems deployed by ICE and CBP have evolved from identity verification tools into part of a broader surveillance infrastructure. The report, The Tech Behind ICE: Oligarchs, Immigration Enforcement and the Threat to Democracy, by Mijente, Just Futures Law, and Surveillance Resistance Lab, argues that the Department of Homeland Security (DHS) is moving biometric and AI-driven enforcement beyond airports, ports of entry, and detention facilities into neighborhoods and immigration operations, putting privacy and civil rights at risk. Its central argument is not about a single biometric system, but the linking of face images, fingerprints, iris scans, and DNA with wider government and commercial data networks used to identify, profile, and target people. The report’s larger argument is that biometrics have become a gateway to a broader surveillance architecture. It says Immigration and Customs Enforcement (ICE) and Customs and Border Protection (CBP) can connect biometric identifiers with immigration records, social media activity, location data, vehicle records, and commercial datasets. A system built ostensibly to locate noncitizens can also reach family members, co-workers, journalists, protesters, legal observers, and others who encounter or document immigration enforcement, and has, the report warns. The authors identify privacy, consent, accuracy, and accountability as their principal concerns. They say federal oversight agencies and civil rights groups have raised questions about privacy violations, improper sharing of photographs, and errors in facial recognition systems. The report argues that facial recognition tools can reveal not only an identity in a public space, but also a person’s professional role, religious affiliation, family, and social relationships, political views, travel patterns, and home address, without consent, a warrant, or probable cause. “Critical civil rights questions remain about the impact on civil, privacy, and consumer rights
Jun 30, 2026 · via biometricupdate.com
It was a full house at the Tuesday, June 23 meeting of the Stoughton City Council, as around a dozen people spoke during the public comment period against the Stoughton Police Department’s use of Flock cameras. Residents spoke of concerns about the use of facial recognition software, violating privacy rights, and the possibility of data being misused by a third party used by the company. In a presentation to the council, Stoughton Police Chief Brandon Hill said the department has been using Automated License Plate Reader (ALPR) cameras since 2019, including a squad car that was formerly equipped with a camera, and asked council members to allow the department to continue using them. He said the department formerly had a contract with a company that was going out of business, so they switched to Flock, a competitor. “We started talking about it in 2018 (and) we got a grant from the Bryant Foundation,” he said. “This isn’t something new. We needed to switch vendors for this tool that we’ve used now for eight years. We found Flock was the best provider in what we were looking for.” How it works The city currently leases cameras from Flock Safety in a contract that expires in December of 2027. They use two types of cameras - ALPRs and pole cameras, which have a live video feed, Cameras were installed in May on major roadways and high police call volume areas. Once construction on Hwy. 51 is completed, SPD will have four ALPR cameras and one pole camera. Once an image of a vehicle or license plate is captured, it is searched through different lists, such as stolen vehicles or missing/endangered people. Alerts notify the department that a vehicle or license plate image was captured and where it was located for officers to
Jun 30, 2026 · via stoughtonnews.com
This editorial was written by Sheena Magenya and Smita V. of the Association for Progressive Communications-Women’s Rights Programme (APC WRP). It is part of Global Voices’ June 2026 Spotlight series, “Gender Diversity.” This series offers insight into gender diversity and how it is being threatened, protected, and preserved around the world. You can support this coverage by donating here. When NASA was preparing for the launch of Space Shuttle Challenger in 1983, the engineers asked Sally Ride, the first American woman in space, if 100 tampons were enough for a six-day flight. In November 2025, the US Transportation Department unveiled THOR-05F, the first crash test dummy specifically based on a woman’s body, and not just a scaled-down version of the crash test dummy based on men’s bodies. Several new model cars today, by default, unlock all doors as soon as the car is put into “park’ or when the ignition is turned off, which can be particularly unsafe for women and LGBTQ+ people. This setting has to be manually changed. Closer to home, in various social justice spaces and conferences, there are often panels and conversations taking place that only have men speaking and moderating. Colloquially called “manels,” they are a visible indicator and reminder of how the lack of gender diversity in a conversation is not seen as a problem or even as something unusual. The literal absence of women and gender diverse people across all levels of design and decision-making can be, and is often, dangerous and life-threatening to over half the global population. Inversely, in many contexts where non-heteronormative sexuality and gender expression are legally and socially criminalized, LGBTQ+ people and women are often scapegoats for moral social decay and are positioned as the reason for failed governance or even natural disasters. Through digital misinformation and disinformation,
Jun 30, 2026 · via globalvoices.org
Proton, the privacy-focused productivity app company, released a public AI chatbot, Lumo, last year. On Tuesday, the chatbot received an upgrade. Lumo 2.0 gives the chatbot a variety of newfound powers, including image recognition and image generation capabilities. Users can now upload pictures into Lumo, then use the chatbot to analyze or edit them. Similar to other LLMs, Lumo can also generate imagery based on a user’s prompt. Version 2.0 also expands Lumo’s capabilities for Projects — the widget that allows users to upload documents and conduct work via Proton’s other products like email and cloud storage. Projects now come with user-controlled persistent memory, which is a function that allows Lumo to recall a user’s preferences across various conversational sessions. Additionally, the company says Lumo’s update makes it significantly more powerful than its previous version. The 2.0 version responds to most queries up to 76% faster than its previous iteration, the company says. The chatbot also comes with a new “thinking mode” for more complex problems or questions. “Lumo 2.0 has been re-engineered from the ground up and the introduction of thinking mode gives it powerful new capabilities,” said Andy Yen, founder and CEO at Proton. “Lumo 2.0 demonstrates that users no longer need to choose between powerful AI capabilities and meaningful privacy protections.” The public version of Lumo appears roughly equivalent to other major chatbots in terms of usefulness. It answers questions in a similar format as Gemini and ChatGPT, with approximately the same level of detail and context. Yet, Proton distinguishes Lumo from other chatbot providers with its privacy protections. It uses what it calls zero-access encryption architecture, which encrypts users’ data in transit and at rest, only allowing access to the user. The company also claims that no server-side logging of sessions is retained, so nobody at
Jun 30, 2026 · via techcrunch.com
You do not have permission to retrieve the URL or link you requested,
If you think this was a mistake please call 1300 134 174 or e-mail customercare@realestate.com.au
and quote the reference number #18.72cfdb17.1782890603.3044377
Jun 30, 2026 · via realestate.com.au
Google announced on Monday that the Gemini app is now offering its personalized Nano Banana-powered image generation feature to a broader audience. Starting today, all eligible users in the U.S. can access the feature for free, a service that was previously only available to Plus, Pro, and Ultra subscribers. Google initially announced that Gemini’s Personal Intelligence feature would get Nano Banana-powered image generation back in April, allowing users to create images that reflect their unique interests. This means that images can be generated based on Gemini’s understanding of your likes and preferences without you having to specify them in your prompt. Gemini utilizes data from your Google account connections — such as Gmail, Google Photos, YouTube, and Search — to achieve this. For example, instead of saying, “Create an illustration of me and my favorite things, such as coffee and baking,” you can simply request, “Create an illustration of me and my favorite things.” Gemini can also pull actual images of you from Google Photos, so you don’t need to manually upload photos. Google initially rolled out the Personal Intelligence feature earlier this year, making it widely available to all U.S. users in March. The company recently expanded this functionality to users in India and Japan. Personal Intelligence is an opt-in feature, allowing you to decide which apps Gemini can access. Once enabled, it is set as the default for every prompt, but you can disable it using a new toggle in the Tools menu. Additionally, last month, Google announced several upcoming updates for the Gemini app, including a new “Daily Brief” feature, a revamped interface, access to AI video model Gemini Omni, and a personal AI agent named Gemini Spark. Notably, Google’s AI chatbot Gemini surpassed 750 million monthly active users (MAUs) earlier this year, reinforcing its position as
Jun 30, 2026 · via techcrunch.com
Proton has announced version 2.0 of its Lumo AI assistant, which the company first introduced last year. This new version ships with a "new architecture" and features like reasoning alongside multimodal image processing and upgraded web search. Lumo 2.0 introduces two reasoning modes: Fast and Thinking, and you can probably guess what each one does. The "Fast" mode prioritizes response speed but may sacrifice deep logical breakdown in the process. On the other hand, the Thinking mode slows down to analyze multi-step problems with a visible thinking state. Previous versions of Lumo were only able to handle text, but version 2.0 changes that with support for image processing. This update brings both image recognition and image generation directly to the assistant. Users can upload a chart, document, screenshot, or photo for the AI to analyze. The system reads this visual data to answer your questions. It also generates custom visuals from text prompts and refines rough sketches into styled images. The assistant can also edit existing photos by swapping backgrounds or removing objects. Other features in this release include upgraded web search capabilities that pull live data directly from the internet. The tool easily retrieves current financial reports and weather forecasts to keep answers accurate. Lumo 2.0 also includes a memory feature to remember stuff you told it in past chats. This allows the assistant to learn your working style and preferences over time. You control this memory completely and can tell Lumo to forget details whenever you want. You can also create your own Custom Lumos, specialized assistants that help you do stuff like writing in a particular style or translating technical terms, while performing specialized research without re-explaining instructions. Proton said that on the Artificial Analysis Intelligence Index, its Lumo 2.0 Lite model scores 127% higher than Lumo
Jun 30, 2026 · via neowin.net
Last summer, Swiss-based Proton launched Lumo, an AI assistant with a privacy-first approach. Today, the company has announced Lumo 2.0, a major update to the chatbot that brings three new features commensurate with its core principles of no logs, no data sharing, and zero-access encryption. Proton says Lumo 2 has been rebuilt on a new architecture that brings its biggest leap in capability to date, with Fast and Thinking modes now available. Fast of course prioritizes speed, while Thinking is optimized for more complex, multi-step reasoning. Proton says Lumo 2 responds to everyday queries up to 76 percent faster than Lumo 1.4. Beyond the new architecture, Lumo 2 also boasts multimodal capabilities such as image generation and image recognition. Users can now upload an image to analyze, create visuals from a prompt or a rough sketch, or edit existing images, all in the same conversation. On top of the new features, Lumo 2.0 has far stronger web search compared to Lumo 1.4, according to the company. There's also a Memory feature that lets Lumo learn your preferences, working style, and ongoing context, with the context window now twice as large. Lumo 2.0 also introduces Custom Lumos, described as enabling purpose-built assistants that can be tailored to specific tasks, such as a research assistant that structures answers the way you need them. Lumo is free to use at Lumo.proton.me and does not require a Proton account when accessed. However, if you have a Proton account, your chat history can be saved using the company's "zero-access" encryption across all your devices. There are also mobile apps for iPhone and Android. For power users, Lumo Plus brings unlimited chats, Projects, advanced image generation, and priority access to the fastest models. Plus costs $12.99 per month, and there's also a Lumo Professional plan for
Jun 30, 2026 · via macrumors.com
AUSTIN, Texas, June 30, 2026 (GLOBE NEWSWIRE) -- Pensa Systems, the industry’s first scalable Vision AI solution for shelf planning and in-store execution, today announced two newly issued, foundational AI patents that address some of the biggest barriers to deploying AI, at scale, in physical retail. Retail stores present a difficult challenge for any AI system. A typical grocery or general merchandising store carries 30,000-50,000 discrete products, many of them visually similar on the shelf. Shelf assortments and layouts are historically planned at headquarters by hand but rarely match what is actually in the aisle. Today, retail associates and brand representatives still maintain and audit shelves largely by hand and by barcode-scanning of individual items one at a time. This activity is tedious and error-prone, but necessary to check for stockouts, misplaced items and to ensure merchandising, promotional displays and pricing are set properly. Early computer vision and AI approaches struggled to keep up with the constant packaging changes and product turnover, producing high inaccuracy rates and the confident-but-wrong outputs now widely known as AI “hallucinations” within Large Language Models (LLMs), and in earlier digital image recognition technologies. The common workaround has been to quietly put remote workers behind the scenes, often referred to as Human in the Loop (HITL), to correct the AI before its data or actions are released. That approach is expensive, still inaccurate, and causes crucial lags precisely where immediate, trustworthy AI would have its largest impact: in the store, during execution. “AI that hallucinates isn’t just wrong, it destroys the trust of those who depend on it,” said Richard Schwartz, President and CEO of Pensa Systems. “Retail is a complex environment. Getting AI right, at scale and exactly when and where it is needed is not optional. The foundational work signaled by these patents is
Jun 30, 2026 · via markets.businessinsider.com
Abstract Cardiovascular complications associated with Post-COVID-19 Patients remain difficult to identify at early stages due to heterogeneous physiological manifestations and the limited integration of imaging and clinical indicators in conventional diagnostic frameworks. To address this challenge, this work proposes a multimodal deep learning framework for cardiovascular disease (CVD) risk prediction by integrating cardiac computed tomography (CT) imaging with structured clinical data. The proposed approach applies Adaptive Bilateral Filtering for image enhancement, Kernel Density Fuzzy C-Means (KDFCM) for myocardial segmentation, and Squeezing Extract Chirplet Transform (SSECT) for extracting multi-scale texture and frequency-based imaging features. Clinical variables are preprocessed through Z-score normalization, interquartile range (IQR)-based outlier removal, and Isolation Forest anomaly detection, followed by Adaptive Starfish Optimization (ASO) for feature selection. Imaging and clinical representations are integrated through feature-level fusion and classified using a Deep Image Recognition–based Generative Adversarial Network (DIR-GAN), where adversarial learning enhances discriminative feature representation and model robustness. Experimental evaluation demonstrates strong predictive performance, achieving 94.98% accuracy, 94.98% sensitivity, 93.38% specificity, 95.43% precision, and 94.74% F1-score. The findings indicate that the proposed framework provides reliable and clinically relevant cardiovascular risk stratification for patients recovering from COVID-19, while demonstrating robustness against heterogeneous multimodal data. Similar content being viewed by others Funding This research received no external funding. Author information Authors and Affiliations Corresponding author Ethics declarations Competing interests The authors declare no competing interests. Ethical approval and informed consent This work did not involve direct interaction with human participants or the collection of new patient data. The imaging dataset was obtained from the publicly accessible COVID‑19 CT Scan Dataset (https://www.kaggle.com/datasets/ahmedtronic/covid-19), while structured clinical records were obtained from the UCI Machine Learning Repository – Heart Disease Dataset (https://archive.ics.uci.edu/dataset/45/heart+disease). All datasets used were publicly available and fully anonymized; therefore, additional ethical approval and informed consent were not required in accordance with applicable
Jun 30, 2026 · via nature.com
The South Korean government will introduce stringent regulatory measures to curb the proliferation of "dae-po" phones—illegally registered mobile devices increasingly linked to voice phishing and financial crimes. The Ministry of Science and ICT (MSIT) announced on June 30 a comprehensive policy framework designed to prevent the illegal use of mobile phones. The measures build upon a government-wide crackdown on telecommunications-enabled fraud initiated last year. According to MSIT, the policy establishes tailored preventive frameworks based on specific types of device misuse, while strengthening post-incident enforcement and legal penalties. The shift reflects the expanding role of smartphones as primary tools for financial transactions and digital identity verification. To prevent identity theft, the government will mandate facial recognition authentication at the point of mobile phone activation. Following a pilot phase conducted last year, the system will be introduced sequentially starting July 6. The finalized implementation plan incorporates guidelines from the Personal Information Protection Commission (PIPC) and the National Human Rights Commission of Korea (NHRCK). During the initial rollout, users opting for facial recognition must complete the verification step within three attempts to proceed with line activation. If facial recognition fails due to technical errors, alternative identity verification methods—such as mobile identification apps or physical identity certificates issued on the same day—will be provided. In October, the government plans to codify the legal grounds for facial recognition by amending the Enforcement Decree of the Telecommunications Business Act. By November, a subscription restriction service, which previously required manual opt-in by users, will be applied by default during contract signing. For foreign nationals, MSIT will collaborate with the Ministry of Justice to upgrade verification infrastructure and strictly enforce a "one person, one line" rule for prepaid SIM cards. The framework also targets "identity lending"—the illegal transfer of mobile lines. To curb fraudulent financial schemes where vulnerable
Jun 30, 2026 · via thelec.net
S. Korea to begin facial verification for new mobile phones next month SEOUL, June 30 (Yonhap) -- South Korea will implement facial recognition technology starting next month as a way to strengthen the identity verification process during mobile phone registrations, the government announced Tuesday. The Ministry of Science and ICT unveiled the plan during a briefing in Seoul as part of broader efforts to prevent illegal activities, including mobile identity theft, when registering new mobile phone numbers. Starting Monday, customers registering a new mobile phone number will be able to choose either face recognition or other government-approved forms of identification, such as resident registration certificates, to verify their identity. Face recognition was made optional, amid concerns over infringements of basic rights, the ministry said. Raw facial images will be discarded immediately after the verification process is completed to prevent potential data leaks, it added. The government plans to expand the range of identity verification options while revising relevant laws by October to establish the legal basis for introducing facial recognition services. Last year, the ministry announced a plan to make face recognition mandatory for new mobile phone registrations, sparking public concerns amid a series of data breaches in local mobile carriers. State data protection and human rights authorities pointed out that the plan lacked a sufficient legal basis and could infringe on people's fundamental rights. fairydust@yna.co.kr (END) - (World Cup) S. Korea eliminated from group stage - At Jeju Forum, U.N. chief candidates vow to reclaim relevance - U.S. nuclear expert calls for patience, caution on Seoul-Washington nuclear submarine cooperation - Security must extend beyond defense in fragmented world: ex-U.S. ambassador - Former leaders call for renewed cooperation amid fragmented global order - At Jeju Forum, U.N. chief candidates vow to reclaim relevance - (World Cup) S. Korea eliminated from
Jun 30, 2026 · via en.yna.co.kr
XAILIENT ANNOUNCES SUCCESSFUL INDEPENDENT VALIDATION OF CASINO EYE-D BY GAMING LABORATORIES INTERNATIONAL (GLI) LAS VEGAS, June 29, 2026 /PRNewswire/ -- Xailient, a leader in Edge AI-powered Facial Recognition (FR) solutions for the gaming industry, today announced that Gaming Laboratories International (GLI) has conducted a successful independent validation of the company's ground breaking Casino Eye-D FR platform. GLI's evaluation confirms the functional and technical specifications of Casino Eye-D, which integrates with the industry's leading casino management systems (CMS) to revolutionize patron identification, loyalty engagement, and AML monitoring across gaming floors. "GLI's validation represents an important milestone for Xailient and the broader gaming industry," said Lars Oleson, CEO of Xailient. "As casinos adopt AI-driven technologies to improve guest experiences and operational efficiency, independent validation provides operators, regulators, and technology partners with full transparency and confidence in these technologies' functionality." Key capabilities reviewed during the evaluation included: - Patron identification and recognition integrated with casino management systems - Automated enrollment and image quality assessment - Enterprise-scale biometric processing and identity management - Support for loyalty engagement and player recognition workflows - Excluded patron identification and security alerting - Secure API-based integration with existing CMS platforms - Scalable deployment architecture designed for casino environments - Confirmed exclusive functionality with CMS and no engagement with any game functionality Doug Beavers, Sr. Director of Business Development said, "GLI's validation provides another step to market deployment of our 'AI with an ROI' facial recognition technology to the gaming floor. We are excited by the support from regulators, our CMS partners, and casino operators as we launch the transformative Casino Eye-D technology to the gaming floor." Xailient has previously announced integrations with Konami Gaming's SYNK Vision™, Light & Wonder CMS and is coordinated with operators, regulators, and industry stakeholders to advance the adoption of AI-powered guest recognition solutions
Jun 30, 2026 · via prnewswire.com
Abstract Simultaneous Localisation and Mapping (SLAM) is a fundamental building block for markerless Augmented Reality systems. However, most conventional SLAM systems exhibit poor performance in real-time environments because of the influence of dynamic objects in unstructured surroundings. We are proposing a new Dynamic Visual Inertial SLAM system called DVI-SLAM to address the challenges of dynamic content presented in complex urban scenes. Leveraging instance-aware segmentation and optical flow methods, our system can detect and exclude actively moving objects from the tracking process. In addition to excluding features of moving objects from the pose estimation process, the system also detects, tracks, and generates static-map of the surrounding environment. This helps develop a robust AR system or mobile robot system to effectively handle complex outdoor urban scene scenarios. The proposed DVI-SLAM framework was tested on the TUM RGB-D and TUM-VI datasets as well as in real-world environments. The results showed an improvement in ATE’s Root Mean Square Error (RMSE) up to 24.80 and 32.34% compared with ORB-SLAM3, on monocular-inertial and stereo-inertial sequences of the TUM-VI dataset, respectively. Similarly, an improvement up to 64.89 and 73.75% was achieved for real-time monocular and RGB-D inertial sequences of real-time environments. Further, the sequence with a lot of dynamic content, such as the TUM-RGBD dynamic dataset, showed up to 99.4% improvement in ATE RMSE. The results demonstrated that DVI-SLAM significantly enhances Absolute Trajectory Error accuracy as compared to the state-of-the-art ORB-SLAM3 method. Extensive experiments evaluation shows that DVI-SLAM performs robustly across both static and dynamic scenarios, while maintaining competitive performance in static and low-dynamic scenes, it delivers significant improvements in highly dynamic environments. Similar content being viewed by others Introduction In Augmented Reality (AR) systems, accurate registration of virtual content with respect to the physical environment is a fundamental requirement for achieving stability and immersive user experience.
Jun 29, 2026 · via nature.com
Abstract Terahertz (THz) imaging technology is widely used in applications such as security screening, radar detection, and biomedical applications. Nevertheless, due to the inherent limitations of imaging conditions, THz images often exhibit low contrast, blurred contours, and reduced feature information under partial occlusion, which significantly compromises recognition accuracy. To address these challenges, we proposed an Enhanced and Occlusion-aware Focus YOLOv5(EOF-YOLOv5), an improved architecture based on YOLOv5. In this study, image enhancement preprocessing was applied to raw THz image datasets acquired from a THz active array imaging system to improve target contrast and contour clarity. An Occlusion Aware Context Attention (OCA) mechanism was integrated into the neck network of YOLOv5. This mechanism dynamically adjusts attention weights and enhances feature responses in visible areas by capturing spatial occlusion patterns via 1 × 1 convolution and modeling inter-channel dependencies. Additionally, the original Complete Intersection over Union (CIoU) loss function was replaced with the Focal-Efficient Intersection over Union (Focal-EIoU) loss function to reduce excessive focus on simple samples and improve the learning performance for challenging samples. Experimental results demonstrate that image enhancement preprocessing significantly improves both visual quality and structural information, thereby boosting the network’s recognition accuracy. On the same preprocessed dataset, the EOF-YOLOv5 algorithm outperforms the baseline YOLOv5 model, elevating precision (P) from 66.7 to 79.3%, recall (R) from 73.1 to 80.6%, and mean average precision (mAP50) from 75.5 to 83.7%. The proposed model effectively identifies targets under blurry and occluded conditions, providing an innovative solution for terahertz image detection. Similar content being viewed by others Funding This work was financially supported in part by the Key Scientific Research Plan of Education Department of Shaanxi [23JY035], The Youth Innovation Team of Shaanxi Universities [K20220184], and Natural Science Foundation of Shaanxi Province [2025JC-YBMS-744]. Author information Authors and Affiliations Corresponding author Ethics declarations Competing interests
Jun 29, 2026 · via nature.com
Can Clothes Make You Invisible to Facial Recognition? Does life feel Orwellian sometimes? One researcher has a solution for you: graphic tees that confuse the neural networks in surveillance cameras. About 10 years ago, an app developer named Hoan Ton-That scraped your social media photos. You didn't know it at the time, but he added your photos to a database, and used it to build a facial recognition platform called "Clearview AI." Your face has been in his database ever since. In the years since, Clearview has only collected billions more photos, attracted investment from plutocrats, and been rewarded with multimillion-dollar contracts from law enforcement agencies across America like Immigration and Customs Enforcement (ICE). First in cautionary science fiction, and then in our timeline, facial recognition has almost always been trained and imposed on regular citizens without their consent. Clearview AI tacitly understood that, and today it's official US government policy. In an analysis of ICE's other facial recognition toy, "Fortify," the Department of Homeland Security (DHS) acknowledged that "ICE does not provide the opportunity for individuals to decline or consent to the collection and use of biometric data/photograph collection." "The surveillance state has overreached," says Bill Swearingen, aka @hevnsnt. "You never opted into this, and there's no way to opt out. That's why I'm trying to give power back to the people a little bit." At Black Hat USA 2026 next month, Swearingen will debut his proposed solution for our totalitarian nightmare: clothing. Graphic hoodies, shirts, scarfs, etc., with patterns that demonstrably confuse or outright break current facial recognition artificial intelligence (AI). The Weak Point in Facial Recognition In the gestalt, facial recognition is an unseen fabric pervading all public spaces. The first step in beating it is making it a smaller, knowable thing. Any given facial recognition flow
Jun 29, 2026 · via darkreading.com
Wearable digital stethoscopes are transforming remote healthcare by enabling continuous monitoring of heart and lung activity outside traditional clinical settings. UNSW researchers developed the AusculPatch, a lightweight wearable sensor that detects subtle mechanical vibrations from the heart, lungs, blood flow, and pulse waves while users go about their daily routines. Unlike conventional wearables that primarily track heart rate or blood oxygen, the patch captures detailed physiological signals and is designed to work with AI systems that can identify abnormal patterns and notify clinicians before symptoms become severe. For healthcare providers, continuous home monitoring could improve chronic disease management while reducing unnecessary hospital visits and enabling earlier intervention. Medical device companies also have opportunities to expand beyond fitness tracking into clinically focused wearable diagnostics that generate richer health insights. As AI-driven analysis becomes more sophisticated, wearable digital stethoscopes could support more personalized care, improve remote patient monitoring, and help healthcare systems manage growing patient populations more efficiently. Image Credit: UNSW Key Themes Behind This Trend - Continuous Acoustic Monitoring - Wearable sensors that capture heart, lung, and vascular sounds create new possibilities for detecting clinical changes during everyday routines rather than brief appointments. - AI-driven Diagnostics - Advanced pattern recognition can turn subtle physiological vibrations into predictive health insights that support earlier intervention and more personalized patient care. - Home-based Chronic Care - Remote monitoring tools expand chronic disease management beyond hospitals and clinics, reducing avoidable visits while giving providers richer longitudinal health data. Where This Applies - Medical Devices - Clinically focused wearables represent a shift from basic fitness tracking toward regulated diagnostic platforms that generate deeper cardiovascular and respiratory intelligence. - Remote Healthcare - Virtual care models benefit from continuous physiological data streams that help clinicians assess patient status without relying solely on scheduled consultations. - Health AI -
Jun 29, 2026 · via trendhunter.com
Can generative AI be an ally in rooting out ransomware threats? A UC researcher suggests rethinking ways to catch bad actors The online publication, Securities.io, reports that ransomware attacks on business activity are projected to exceed $265 billion annually by 2031. Using generative AI to find a way to help alleviate these cybersecurity threats may be a smart option. Securities.io cites recent research published by Nelly Elsayed, associate professor in the UC School of Information Technology, in the Journal of Information Security and Applications, which suggests that generative AI may be an ally in strengthening ransomware defense. “We are in a hype era of AI,” says Elsayed, founder and leader of the Applied Machine Learning and Intelligence Lab within the College of Education, Criminal Justice, and Human Services. “Some people support it, others fear it, but in general people who design technology are trying to use it for good.” Elsayed's research article argues that generative AI can be used to integrate synthetic data generation and behavioral forecasting, stress test systems by checking for adversarial behavior simulation and improve trust of human-AI collaboration in security operation systems. Cybersecurity analysts and system defenders can use AI to detect new malicious attacks and classify and identify new means of attack from bad actors, according to Elsayed. Simulating with hackers might allow for creating possible attack scenarios and learning to think like attackers to offer more robust tools for defense, she adds. “It’s a way to generate a combination of possible attacks system defenders might not have considered,” she says. Elsayed adds a practical example could be a user pasting a suspicious email into a generative AI system and asking about the validity of the email. AI could help screen and catch red flags: a suspicious logo or misspellings. “AI can become an early
Jun 29, 2026 · via uc.edu