Wrongfully Arrested Peppermill Truck Driver Slams Facial Recognition Tech Posted on: April 8, 2026, 03:49h. Last updated on: April 9, 2026, 06:14h. Lawyers acting for a man wrongly arrested at Reno’s Peppermill Casino following misuse of AI-based facial recognition say poor training has led police to unlawfully detain “thousands” in similar cases. In September 2023, truck driver Jason Killinger was flagged as a “100% match” to an individual named Michael Ellis who had been banned from the venue months earlier for sleeping on the premises. Held for 11 Hours Killinger was detained by casino security and then arrested by Richard Jager, a rookie officer from the Reno Police Department. Jager refused to believe Killinger was not who he said he was and accused him of obtaining a fraudulent ID when he provided evidence of his identity. He was detained for 11 hours, and handcuffed for four, which resulted in bruises and shoulder pain, according to a lawsuit he later filed against Jager. The suit claims that even after his real identity was confirmed by a fingerprint check at Washoe County jail, Jager filed a police report saying that Killinger had given conflicting IDs to Peppermill security. City Added to Lawsuit Last week, federal Judge Miranda Du granted a motion to add the City of Reno to the lawsuit as a defendant. The primary reason was the city’s continued pursuit of legal action against Killinger after he had been exonerated, and because it allegedly failed to properly train police officers on the limits of facial recognition technology at the time of Killinger’s arrest. However, a new filing from Killinger’s lawyers alleges that Reno police have long been aware that facial recognition results aren’t reliable enough to justify an arrest, but the department continued to follow that approach under Chief Kathryn Nance.
Apr 9, 2026 · via casino.org
Adversarial Examples in Computer Vision: How Attacks Work and How to Build Robust Models Adversarial examples in computer vision are inputs that appear normal to humans but cause neural networks to make confident, incorrect predictions. What began as small, gradient-based pixel tweaks has expanded into physically realizable attacks (patches, textures, camouflages) and latent-space manipulations that target internal representations. By early 2026, research has increasingly focused on vision foundation models and vision-language systems, where multimodal attack surfaces such as prompt injection and jailbreaking have become practical concerns. This article explains how adversarial attacks work, why they transfer across models, how physical attacks succeed outside the lab, and which defenses actually improve robustness in real deployments. Adversarial attacks exploit model sensitivity to input perturbations, impacting vision systems and inference reliability-build defensive expertise with an AI Security Certification, implement robust CV pipelines using a Python Course, and align model outputs with real-world applications through an AI powered marketing course. What Are Adversarial Examples in Computer Vision? Adversarial examples are intentionally crafted perturbations applied to an image that cause a machine learning model to misclassify it. The defining property is that the perturbation is often imperceptible or resembles benign noise, yet reliably changes the model output. In computer vision, this can affect: Image classification (mislabeling an object) Object detection (missing an object or hallucinating one) Segmentation (corrupting pixel-level masks) Biometrics (face recognition evasion or impersonation) Autonomous driving perception (sign detection, vehicle detection, sensor fusion) Recent surveys underscore a dual reality: adversarial examples represent a security threat and also serve as a testing tool for building more resilient models. Many modern defense strategies explicitly reuse attack methods to harden systems. How Adversarial Attacks Work Most adversarial attacks exploit how deep neural networks respond to small input changes in high-dimensional spaces. Even a tiny change in
Apr 9, 2026 · via blockchain-council.org
Lip reading, emotion sensing, face biometrics vie for place in smart auto stack Your car may soon know when you’re tipsy, stressed, or just mouthing words into the wind. From Ford’s lip-reading patent to radar systems that sense your mood and (of course) biometric driver identification, the latest wave of automotive AI is turning vehicles into surprisingly perceptive co-pilots. Ford publishes patent for in-car lip reading Vehicle voice control systems have been around for a while – but soon, they may be able to listen in on our most secretive conversations. Ford has filed a patent application for software that reads the occupant’s lips using cameras and sensors inside the vehicle. The feature would not be used for lip-based biometric authentication. Instead, the automaker says the vehicle would turn on lip reading in situations where normal voice controls do not work due to wind or other noise, for example, while driving in a convertible with the top down. Aside from lip movement captured by cameras, the system could also rely on acoustic signals: the vehicle could emit inaudible sound waves and analyze the echoes bouncing back from the user’s lips and mouth with the help of machine learning. “The captured video and sensor data are processed using machine learning algorithms that are trained on large datasets of lip movements and corresponding speech to learn the patterns and nuances of lip reading,” says patent 20260095520, recently published by the USPTO. The patent also predicts the use of gesture- and facial-expression-detection software to determine whether the user is having difficulty interacting with the system. Nodding the head could be interpreted as acknowledging a verbal interaction with the vehicle, while shaking the head or appearing confused would lead the system to believe that the user is having difficulties. The vehicle could then choose
Apr 9, 2026 · via biometricupdate.com
PRAGUE, CZECH REPUBLIC - Apr 8, 2026 - Future of Gaming, the strategic intelligence platform tracking gaming patents from the USPTO, has published its Q1 2026 Quarterly Granted Report analyzing 124 patents awarded to 21 companies. The report reveals a consistent industry priority: eliminating player friction through automation, from controller drift calibration to backend service integration to crafting systems that don't pull players out of the game world. "The granted patents tell you what's actually shipping," said Alex Kirillov, founder of Future of Gaming. "Filed patents show intent. Granted patents show commitment. When Sony gets 27 grants in a single quarter spanning everything from gesture recognition to spatial anchoring to procedural animation, that's not speculative R&D. That's infrastructure for products in development right now." Key Findings Sony received more than triple the next-largest company. Sony's 27 granted patents span six categories, with 16 in AI/ML alone. Their grants cover gesture recognition that prevents false positives during gameplay, spectator-driven AI adjustment, session duration prediction, customizable flat controllers for accessibility, and spatial anchoring for mixed-reality experiences. Cross-platform led granted patents with 52. Unified progression tracking, backend service integration, and multi-game loyalty systems dominated. Betty Gaming received a patent for universal progress bars converting game-specific metrics into platform-wide rewards. Amazon patented pre-configured workflows connecting authentication, matchmaking, leaderboards, and voice chat without custom integration code. AI/ML accounted for 31 grants. EA's Motion Variational Autoencoders generate character poses in real-time without storing animation data. Cygames patented automated bug detection for collectible card games by comparing irregular and regular gameplay session logs. Colopl received a patent for AI-generated NPC dialogue replacing scripted responses. Hardware grants focused on accessibility and durability. Sony received patents for adaptive dead-zone calibration, touch-sensitive control surfaces, and customizable flat controllers. Wi-Charge patented laser-based wireless charging for controllers during active gameplay. Backbone Labs
Apr 8, 2026 · via markets.financialcontent.com
ICE Already Knew Her Name The Tech: Mobile Fortify ICE agents are using a phone app called Mobile Fortify. Mobile Fortify allows agents to: - Take your photo; - Scan your face and fingerprints; - Instantly search state and federal government databases. With Mobile Fortify, agents simply need to point a phone at anyone in public in order to compare their face against databases containing 200 million images and instantly access their name, date of birth, and other data. You May Already Be In the Network Hundreds of millions of people are in these systems, including U.S. citizens here in Wisconsin. You don’t need to have previously interacted with ICE to already be in the database. Mobile Fortify connects to databases including: - Federal biometric systems; - Passport and visa records; - State driver’s license photos. If you do happen to enter the database through a face scan in the app during an interaction with federal agents, the government keeps the scan up to 15 years regardless of immigration status. In comparison, TSA claims to delete your facial recognition data at airports after 24 hours if you are a U.S. citizen. Error-Prone Privacy Nightmare Facial recognition doesn’t confirm identity — it guesses. False matches happen. One person can be matched to multiple different identities (including wrongly flagging U.S. citizens as undocumented immigrants). Furthermore, there are higher error rates for Black and Brown people. Congressman Bennie Thompson, ranking member of the House Homeland Security Committee, stated that ICE indicated that it treats Mobile Fortify as a “definitive” determination of a person’s status and that an ICE officer may ignore evidence of U.S. citizenship, including a birth certificate, when the app says a person is undocumented. Technology that is unreliable even in controlled settings should not be used to doll out consequences
Apr 8, 2026 · via aclu-wi.org
Shoplifting and thefts down 10%, says police force Shoplifting and theft offences in Cheshire have fallen by 10% during the last 12 months, police have said. According to Cheshire Constabulary, this means the county has been bucking the national trend. Assistant Chief Constable Alison Ross said her force was working more closely with retailers and local authorities, as well as making good use of CCTV footage and facial recognition technology. "All the different technologies we now have to support businesses, not only to gather the evidence, but also identify the suspect, is incredibly important," she said. Ross urged shop owners to continue to work with Cheshire Police to ensure their CCTV cameras are working properly, and to effectively share footage of any suspicious activity. "Wherever you are in the community - whether that's a large business or a small business - we can give you advice relevant to your particular business," said Ross. Carl Critchlow, chief executive of the Chester Business Improvement District, said while businesses had reported a reduction in shoplifting over the last six months, it remained a significant issue. He said he wondered whether "we're seeing less shoplifting or whether we're just getting better detecting it". Critchlow added: "I think that's the challenge and I think that's what we're working with - our businesses and the police - to try and understand. "We are seeing improvements, but there is still probably a lot of work that we need to do in terms of reporting to make sure that all incidents do get reported and the police have a true understanding and a real picture of what the issue is." Sophie Bryant, who runs The Gift Box in the Hoole area of Chester, said she continued to have problems with shoplifters. She said police were inconsistent in the
Apr 8, 2026 · via bbc.com
What began as low-budget, authentic programming evolved into a cable TV industry capable of producing prestige content that redefined media and entertainment. Are today’s creators traversing the same trajectory? The parallels are striking, but so are the differences, and those differences present both potential limitations and inherent advantages as the creator economy continues to evolve and grow. Cable TV’s Evolution: From Raw Authenticity to Prestige Storytelling in Three Phases From the 1990s into the 2020s, the cable TV industry* delivered ever-increasing levels of fidelity in content and achieved greater and greater heights, cementing itself as a backbone of modern media. This evolution unfolded across three distinct phases, each representing a leap in storytelling ambition and production value. With each phase, cable’s audience followed and the industry thrived. Phase 1 – Unscripted Reality: First-Person Storytelling + Low-to-Mid Fidelity – Cable first captured audiences through reality content built on first-person, unscripted authenticity. In first-person storytelling, stories unfold through an individual’s direct perspective and lived experience. The audience participates from within rather than observing from the outside. This immediacy is the core of the parasocial bond that makes reality content so compelling. In the 1990s, cable pioneered this format with shows such as MTV’s The Real World (1992 – 2017) and Road Rules (1995 – 2007), which later extended into Laguna Beach (2004 – 2006), Deadliest Catch (2005 – present), The Real Housewives (2006 – present), Keeping Up With The Kardashians (2007 – 2021) and many others, giving viewers unfiltered access to “real people” in niche communities. This Phase 1 content is cost-effective to produce and sustains engagement across long seasons and/or reruns, building deeply engaged fan bases. Phase 2 – Scripted Animation: Third-Person Storytelling + Low-to-Mid Fidelity/Budget – Soon after reality content, low-to-mid-budget animation found a home on cable. Cable’s Phase 2
Apr 8, 2026 · via jdsupra.com
China’s draft rules on AI ‘virtual humans’ target biometric deepfakes China has issued a new draft regulation on “digital virtual humans,” or AI chatbots, introducing clear labeling requirements and mandating explicit consent for using a person’s likeness, voice, or other personal data to create a virtual representation. The regulation would also ban the use of digital humans to “evade facial recognition, voice recognition, or other identity authentication mechanisms.” The Cyberspace Administration of China (CAC) published the draft rules, titled Measures for the Management of Digital Virtual Human Information Services, last Friday. Public consultations will run until May 6th. Digital virtual humans are defined as virtual figures that exist in nonphysical environments and simulate human appearance and behavior using technologies such as computer graphics, digital image processing, and artificial intelligence. In other words, personas used by chatbots based on LLMs like ChatGPT and Grok, and the myriad applications based on them. The clauses on consent for use of likeness and banning bypass attacks against biometric authentication systems clearly targets the spread of deepfakes. The rulebook introduces special protections for minors, prohibiting the offering of virtual intimate relationships to children under 18, including simulated family members or romantic partners, as well as any services that may trigger excessive spending, promote harmful conduct, or compromise their physical or mental well-being. The draft also introduces measures to prevent the spread of rumors, insults and defamation through digital humans. Organizations and individuals should resist the creation or distribution of content containing sexual innuendo, cruelty, horror, or discriminatory material, according to the draft (translation). The regulation is part of China’s efforts to protect personal data and rein in AI technologies. The draft law builds on CAC’s 2025 Measures for Labeling of AI-Generated Synthetic Content, which require visible and technical labels for AI-generated text, images, audio and
Apr 8, 2026 · via biometricupdate.com
Sponsor Content Created With Dell Technologies Top business tips for the Dell Pro 5 and Dell Pro 7 The Dell Pro 5 and Dell Pro 7 unlock a slew of benefits for businesses So, you’ve picked the Dell Pro 5 and Dell Pro 7 for your workforce. These devices are the latest and greatest in the Dell Pro lineup, providing all the power needed for today’s business software as well as sleek, professional designs that suit any table – be it boardroom or train. For all these reasons and more, leaders who land on the Dell Pro 5 or the Dell Pro 7 for their workforce device upgrade can rest easy knowing they’ve chosen some of the best business devices on the market. To really get the most out of the investment, though, it’s worth knowing the top business tips for the Dell Pro 5 and Dell Pro 7. TL;DR: - One of the primary benefits of upgrading to the Dell Pro 5 or Dell Pro 7 is to make the most of Microsoft Copilot on Windows 11, as well as other AI apps and features that require an NPU to work properly. - Wherever you're headed, be it the office commute or a trans-oceanic flight, the Dell Pro 5 and Dell Pro 7 can meet your needs and provide the battery needed to make the journey with you. - Dell Pro laptops are equipped with top of the line displays, sophisticated cameras, and clear microphones, to remove friction between a user and all the people with whom they interact. Make the most of AI apps Both the Dell Pro 5 and Dell Pro 7 laptops provide ample opportunity for staff to leverage AI applications and features, including those that come as a part of Windows 11. This is because
Apr 8, 2026 · via itpro.com
Man sues City of Reno over use of facial recognition
RENO, Nev. (KOLO) - A man claiming to have been wrongfully arrested at the Peppermill is now suing the City of Reno.
According to court documents, back in 2023, Jason Killinger was at the Peppermill when facial recognition software identified him as a man who was earlier barred from the property.
Killinger was arrested and taken to the Washoe County Detention Facility for fingerprint testing to confirm his identity. Killinger was determined to not be the man in question and he was released.
In 2025, he sued the arresting officer, alleging fabrication of evidence and excessive force. The court denied the claim of wrongful arrest.
In a new filing, a district court has granted his request to add the City of Reno as a new defendant in the case, as well as a new municipal liability claim against the city.
Copyright 2026 KOLO. All rights reserved.
Apr 8, 2026 · via kolotv.com
Startup launches deepfake detection capable of tracing images to specific tools Canada-based content analysis provider Winston AI has launched a forensic image detector that the company says can trace deepfakes back to their source. The new tool can pinpoint which parts of an image were manipulated, according to Winston AI, as it can determine how they were altered, and which AI model or editing tool produced the changes. “Until now, detection tools told you whether to trust an image,” says John Renaud, CEO of Winston AI. “Ours tells you exactly where not to trust it, and which tool put it there. That’s the difference between a smoke alarm and a fire investigator.” More than 10 million people use the AI content-detection platform, which Winston AI designed for high stakes environments such as journalism and fact checking, legal evidence review, academic integrity checks and content moderation teams. In principle, it could also have value for fraud prevention teams handling images submitted for KYC checks that they suspect may be biometric deepfakes. Instead of offering a binary “real‑or‑AI” verdict, the system analyzes images using six independent forensic techniques. These map pixel‑level anomalies such as broken camera‑sensor fingerprints, inconsistent compression, unnatural edges and noise patterns. When combined with metadata and C2PA provenance signals, Winston AI’s tool is, when possible, capable of attributing an image to a specific generative model or editing software. A multimodal reasoning model compiles the findings into a structured report that includes an authenticity verdict and confidence score. The software produces a written summary along with region‑by‑region manipulation breakdowns and tool attribution where possible. The forensic image detector is available now on the company’s website in both a lightweight screening mode and a full advanced forensic mode. Deepfake detection is an emerging area. Biometric Update and Goode Intelligence have a
Apr 8, 2026 · via biometricupdate.com
Introduction Facial Recognition Statistics: Facial recognition technology is becoming a part of our daily lives. It helps identify people by studying their faces. This system makes security faster, smartphones smarter, and online services more personalized. From unlocking devices for use in airports to its growing role, it’s a fast-changing area. However, this technology also brings concerns. There are questions about privacy, safety, and possible mistakes. It is important to understand how it works, along with its benefits and risks, as it becomes more common in everyday life. Editor’s Choice - In 2026, the total market size is projected to be about USD 9 billion. - The software segment is projected to hold a 53.9% share by the end of 2026. - In terms of technology, 2D facial recognition is projected to lead, accounting for 41.7%. - By application, security and surveillance will dominate, accounting for 39.4% in 2026. - North America is expected to remain the largest market, holding 43.1% share. - Ayonix Corporation leads the market with a 15% share, followed by IBM Corporation at 11% and Aware, Inc. at 10%. - About 68% of users apply facial recognition to unlock their phones, laptops, or personal computers. - In 2026, almost 55% of Americans are most comfortable with the use of FRT in airports for safety purposes. - The acceptance of FRT is highest for passport control at 81%, followed by building access at 72%. - FRT is widely accepted for security purposes, with 81% approval for passport control and 72% for building access, while 55% of people support its use in police surveillance. - About 60% of adults are somewhat aware that employers use facial recognition technology. - Adoption is higher among younger users, with 12% of users aged 18-34 using it in shared spaces versus 1% of
Apr 8, 2026 · via bayelsawatch.com
Abstract In-sensor neuromorphic vision platforms with near-infrared (NIR) sensitivity are essential for intelligent imaging in low-light and multispectral environments. Here, a retinocortical dual-mode platform based on evolved-synaptic transistor (Evo-SynT) devices using upconversion nanoparticles (UCNP) and their integration into evolved-retina optical synapse (EROS) arrays is introduced. Evo-SynT devices exhibit key synaptic features, including paired-pulse facilitation indices exceeding 183.93% and 136.36% at a 0.5 s interval under 808 nm and 940 nm illumination, respectively, and analog weight modulation across 512 conductance states. A 12×12 EROS array enables dual-mode operation: retinal-like in-sensor preprocessing and cortical-like in-memory classification. The EROS array improves pedestrian detection accuracy under low-light conditions from 0.7806 to 0.8481 (808 nm) and 0.9071 (940 nm), and achieves classification accuracies of 77.19% and 79.40%, respectively. These results highlight the EROS platform as a scalable solution for integrated NIR-sensitive neuromorphic vision systems. Data availability The data supporting the findings of this study are available within the paper and its Supplementary Information files. The source data underlying the graphs and plots presented in the figures are provided with this paper. Additional data related to this study are available from the corresponding authors upon request. Requests for data will typically be responded to within two weeks, and the data will remain available for at least three years following publication. Source data are provided in this paper. Code availability All custom codes used for data processing, analysis, and simulation in this study are provided with this paper as Source Code files. References Huang, P.-Y. et al. Neuro-inspired optical sensor array for high-accuracy static image recognition and dynamic trace extraction. Nat. Commun. 14, 6736 (2023). Zhou, G. et al. Full hardware implementation of neuromorphic visual system based on multimodal optoelectronic resistive memory arrays for versatile image processing. Nat. Commun. 14, 8489 (2023). Huang, H. et al. In-sensor
Apr 8, 2026 · via nature.com
What began as low-budget, authentic programming evolved into a cable TV industry capable of producing prestige content that redefined media and entertainment. Are today’s creators traversing the same trajectory? The parallels are striking, but so are the differences, and those differences present both potential limitations and inherent advantages as the creator economy continues to evolve and grow. Cable TV’s Evolution: From Raw Authenticity to Prestige Storytelling in Three Phases From the 1990s into the 2020s, the cable TV industry* delivered ever-increasing levels of fidelity in content and achieved greater and greater heights, cementing itself as a backbone of modern media. This evolution unfolded across three distinct phases, each representing a leap in storytelling ambition and production value. With each phase, cable’s audience followed and the industry thrived. Phase 1 - Unscripted Reality: First-Person Storytelling + Low-to-Mid Fidelity - Cable first captured audiences through reality content built on first-person, unscripted authenticity. In first-person storytelling, stories unfold through an individual’s direct perspective and lived experience. The audience participates from within rather than observing from the outside. This immediacy is the core of the parasocial bond that makes reality content so compelling. In the 1990s, cable pioneered this format with shows such as MTV’s The Real World (1992 - 2017) and Road Rules (1995 - 2007), which later extended into Laguna Beach (2004 - 2006), Deadliest Catch (2005 - present), The Real Housewives (2006 - present), Keeping Up With The Kardashians (2007 - 2021) and many others, giving viewers unfiltered access to “real people” in niche communities. This Phase 1 content is cost-effective to produce and sustains engagement across long seasons and/or reruns, building deeply engaged fan bases. Phase 2 - Scripted Animation: Third-Person Storytelling + Low-to-Mid Fidelity/Budget - Soon after reality content, low-to-mid-budget animation found a home on cable. Cable’s Phase 2
Apr 8, 2026 · via lexology.com
Abstract DICOM image pseudonymization is an effective measure to protect patient privacy and ensure compliance with the GDPR. However, no standardized methods guarantee the irreversible anonymization of DICOM images or provide evidence on the robustness of these procedures. Organizations such as NEMA propose pseudonymization profiles for DICOM metadata, but the risk to data protection is assumed by the entity responsible for pseudonymization, as the potential privacy risks associated with their use cannot be accurately assessed. In the Re-identification Challenge, the selected 68 participants were tasked with re-identifying 38 pseudonymized DICOM studies from multiple European acquisition sites, modalities and anatomical regions. The data were pseudonymized using the pseudonymization profiles developed in ChAImeleon1 and ProCancer-I2 European AI4HI projects. No participants were able to trace back the identity of the patients, demonstrating the effectiveness of these pseudonymization profiles. Despite minor vulnerabilities identified in areas such as free-text metadata, 3D reconstructions, and usability limitations, no successful re-identification occurred—even with monetary rewards offered to participants. Similar content being viewed by others Data availability The dataset used in this study is publicly available in Zenodo with https://doi.org/10.5281/zenodo.170346748. Code availability Several software tools were used during the Re-identification Challenge. The dataset8 was pseudonymized using the CTP app (version released on 2023-11-04) with project-specific configurations. The CTP app is available in https://github.com/johnperry/CTP/raw/master/products/CTP-installer.jar and its documentation in https://mircwiki.rsna.org/index.php?title=MIRC_CTP. Moreover, the software used to apply de-facing in the dataset8 is available in https://www.nitrc.org/projects/mri_reface. References Bonmatí, L. M. et al. CHAIMELEON Project: Creation of a Pan-European Repository of Health Imaging Data for the Development of AI-Powered Cancer Management Tools. Front. Oncol. 12, 742701. https://doi.org/10.3389/fonc.2022.742701 (2022). ProCancer-I Project. An AI Platform integrating imaging data and health records for precision medicine in cancer management. https://www.procancer-i.eu/ (2025). Regulation (EU) 2016/679 of the European Parliament and of the Council of 27 April 2016 on the
Apr 8, 2026 · via nature.com
Abstract To achieve object detection in UAV vision, an object detector is crucial, but haze seriously affects detector performance by physically suppressing high-frequency information, which makes it difficult to detect tiny objects. Traditional “dehazing-then-detection” paradigms are restricted to inconsistencies in tasks and restoration effects. The article provides Frequency-Domain Modulation Network (FDMNet), which is an aerial non-explicit-restoration haze-resistant object detector. Following a physical prior, FDMNet constructs an estimate-compensate architecture. A Frequency Domain Modulation (FDM) module is one that expressly estimates the loss of spectral information, and a dynamic LMF-Kernel adaptively restores the loss of mid- and high-frequency discriminative information. To be effective, we present a physical-semantic consistency loss strategy, with frequency-domain consistency loss guaranteeing physical accuracy and prompt distillation loss guaranteeing semantic consistency. We do not have enough datasets, and, therefore, we build Hazy-DOTA, Hazy-DroneVehicle, and a real-life UAV test set. Through extensive experimentation, FDMNet achieved a 7.9% improvement in mAP scores on the HazyDet dataset compared to baseline models, alongside respective gains of 9.9% and 11.5% on the Hazy-DOTA and Hazy-DroneVehicle datasets. Furthermore, it attained state-of-the-art performance relative to several advanced algorithms, balancing both accuracy and robustness. Data availability https://www.kaggle.com/datasets/xiaoxiongzhou/hazy-datasetsThe datasets used and/or analysed during the current study available from the corresponding author on reasonable request. References Wang, H. Y. et al. Cross-Modal Oriented Object Detection of UAV Aerial Images Based on Image Feature. IEEE Trans. Geosci. Remote Sens. 62, 1–21 (2024). Xie, X. X. et al. Fewer is more: efficient object detection in large aerial images. Sci. China-Information Sci. 67 (1), 112106 (2024). Hou, T. et al. MFEL-YOLO for small object detection in UAV aerial images. Expert Syst. Appl. 291, 128459 (2025). Tan, Z. W., Jiang, Z. G., Guo, C. & Zhang, H. P. WSODet: A Weakly Supervised Oriented Detector for Aerial Object Detection. IEEE Trans. Geosci. Remote Sens.
Apr 8, 2026 · via nature.com
Tech Trends: New Tools for Venue Management, Facial Authentication, Ticketing and More TSNN takes a deep dive into the newest tech solutions for events of all types, plus the latest must-know industry news. April 8, 2026 Event Technology We're Excited About This Month To give every attendee a personalized, AI-powered event guide Event experience platform Bizzabo has introduced BizzyAI, an attendee copilot designed to deliver real-time, personalized guidance throughout the event experience. Built into the Bizzabo mobile app, Bizzy answers questions, recommends sessions and connections, and helps attendees navigate agendas and venues based on live event data—reducing the need for support staff while ensuring consistent, up-to-date information. “Attendees are more selective about where they spend their time. When they commit, they expect relevance, personalization, and clear outcomes,” said Alon Alroy, co-founder and CMO of Bizzabo. “Bizzy helps meet those expectations by delivering timely, personalized guidance throughout the event experience.” To visualize and plan events in real venues before arriving on site Event production company Freeman has launched Blue Echo, an AI-powered 3D visualization platform that turns convention centers into photorealistic, fully navigable digital environments for event planning. By creating accurate, immersive models of venues, the tool allows organizers to test layouts, map attendee journeys, and evaluate sponsorship placements months in advance—helping teams reduce planning time, avoid on-site surprises, and make more informed design and budget decisions. “Innovation is our instinct,” said Janet Dell, CEO of The Freeman Company. “At TFC, we don’t wait for the industry to change; we build what changes it. Freeman Blue Echo reflects our commitment to intelligent innovation: emerging technology with real, practical value for our clients and our teams.” Photo Courtesy of Freeman To manage venue operations and trade shows, all in one place Venue management platform Thynk and exhibition management platform ShowCycle have partnered
Apr 8, 2026 · via tsnn.com
Fargo mayoral candidate Josh Boschee addresses billion-dollar debt, facial recognition controversy FARGO, N.D. (Valley News Live) - North Dakota State Senator Josh Boschee sat down with Valley News Live to discuss his campaign for Fargo’s first full-time mayor position, addressing the city’s billion-dollar debt, a facial recognition case and a proposed flavored nicotine ban which were among numerous topics discussed. Boschee said he views the city’s debt similarly to a mortgage. According to Boschee, about $550 million relates to infrastructure including water, roads and wastewater as the city grows. Another $200 million involves the wastewater treatment plant, and close to $200 million relates to water intake processes and clean water. “Certainly a billion dollars is a large number. I’m used to working with a $20 billion budget out in Bismarck,” Boschee said. Boschee said the city should apologize for its role in the Angela Lipps case. Lipps, a Tennessee grandmother, spent months behind bars after Fargo Police used facial recognition technology that wrongly identified her in a bank fraud case. Charges were dismissed, and the city has not apologized. “Well I think first and foremost better communication again from the city or law enforcement up front would have helped alleviate a lot of concerns the public had,” Boschee said. “Should Fargo apologize? I think Fargo should apologize for the role that we had.” Boschee said he was disappointed the City Commission tabled discussion on a proposed flavored nicotine ban until after the June election. He said the commission is setting itself up to be a lame duck commission. “They still have two months worth of work, three months actually until the new commission takes place,” Boschee said. Boschee said he wants to visit with all stakeholders before any proposal moves forward, including public health advocates, young people and store owners.
Apr 8, 2026 · via valleynewslive.com
Google vs. Ring vs. Arlo: We put 6 security cameras to the test to find the most accurate AI There's a clear winner Here at Tom’s Guide our expert editors are committed to bringing you the best news, reviews and guides to help you stay informed and ahead of the curve! You are now subscribed Your newsletter sign-up was successful Want to add more newsletters? Join the club Get full access to premium articles, exclusive features and a growing list of member rewards. Many of the best home security cameras on the market today call themselves "smart," but there's a massive difference between a camera that fires off alerts every time a tree branch moves and one that can actually tell you what's happening at your home. Cameras these days use AI to better analyze motion and classify what they're seeing or hearing, and many of them are leaning into generative AI for things like summaries and descriptions, so you can better wade through the hours of footage they record. Of course, AI has become a bit of a buzzword, so I wanted to compare several cameras to see if they actually delivered on the promise of using AI to improve home security. To do so, I tested cameras from six major smart home companies — Google Nest, Arlo, Wyze, Ring, Blink, and Eufy — to see how they use AI, and if it’s worth the extra expense. Article continues belowThe contenders We deliberately picked models from each brand that carry the company's most capable AI feature set. In other words, while we didn't necessarily use the most expensive model that each manufacturer has to offer, we specifically made sure that it had the same AI features. To be sure, many companies' AI features work off-camera, so the particular type
Apr 8, 2026 · via tomsguide.com
Wide-area search 2.0: Strong increase in police facial recognition AI upgrade and mobile apps are causing BKA facial image query numbers to rise sharply. Critics warn of digital stigmatization of entire groups. Digital surveillance in public spaces is reaching a new dimension. Figures from the Federal Ministry of the Interior show a significant increase in the use of the police facial recognition system (GES). Last year, German authorities used the system, which is part of the Federal Criminal Police Office (BKA), much more frequently for identifying people than before. With a total of around 343,856 searches in 2025, the frequency more than doubled compared to the previous year. According to a response from Interior State Secretary Christoph de Vries (CDU) to a request from Bundestag member Lea Reisner (Die Linke), the criminal police offices of the federal and state governments stand out: in 2024, they still generated around 121,000 queries. But recently, the figure shot up to over 313,500 searches in 2025. The Federal Police also intensified its use, accessing the central photo database about 30,000 times – an increase of around 50 percent. Established in 2008, the system officially serves as an assistive tool to clarify the identity of suspects or victims. The procedure is based on comparing image material from surveillance cameras or mobile phones with the Inpol file. This currently contains 7.6 million photographs of approximately 5.4 million people. The software encodes anatomical features of a face into a mathematical template and compares it with the database in a fraction of a second. The result is a list of candidates sorted by probability. Results must subsequently be manually verified by photo experts. AI use and mobile investigation via app The technical modernization is the driving force. Since September 2024, the BKA has been using a system based
Apr 7, 2026 · via heise.de