As image processing becomes central to industries from manufacturing to healthcare, organizations face a crucial question: should visual data be processed in the cloud or at the edge?

The explosion of high-resolution video from smartphones, IoT devices, and industrial cameras is driving demand for fast, efficient processing. In latency-sensitive scenarios like autonomous vehicles or real-time surveillance, delays can be costly — even dangerous. At the same time, privacy concerns and rising cloud costs make data locality and bandwidth efficiency top priorities.

• Cloud or SaaS computing offers scalable power and easy access to advanced AI models — ideal for heavy processing and large-scale tasks.
• Edge or Embedded computing, on the other hand, brings speed and control, enabling real-time analysis and better data privacy by processing directly on the device.

The choice between cloud and edge isn’t just technical — it’s strategic. In many cases, hybrid approaches that combine both will define the future of AI-powered vision systems.

​Why SaaS-Based Computer Vision Is Gaining Traction

As AI adoption expands across industries, many organizations are turning to SaaS-based computer vision platforms to unlock the power of visual intelligence — without the overhead of managing infrastructure or models. While edge AI remains critical for real-time and mission-critical systems, Software-as-a-Service models offer unique advantages, particularly for web-centric and analytics-heavy workflows.

Efficiency Through the Cloud

SaaS computer vision removes the complexity of deploying and managing AI systems. Instead of setting up GPUs, storage, or embedded devices, companies can access ready-to-use tools that process images, videos, and streams via simple APIs or dashboards. This makes rapid prototyping and scalable automation accessible to startups and enterprises alike.

Businesses can go from raw data to insight in minutes — without touching a single server.

Benefits of the SaaS Model

• Speed to deployment: No hardware, no setup. Users simply connect their cameras or upload data and begin analysis.
• Elastic scalability: Cloud platforms handle millions of frames or photos daily, scaling automatically with demand.
• Automatic updates: AI models are continuously improved by the provider, ensuring access to the latest features without manual upgrades.
• Ease of integration: REST APIs, SDKs, and low-code options make it easy to connect SaaS vision tools to existing systems — from CRMs to warehouse software.
• Cost efficiency: For post-processing, non-real-time use cases (e.g., defect classification, visual search, document recognition), SaaS platforms are typically more economical than edge or hybrid solutions.

But Not Always the Right Fit

Despite its strengths, SaaS computer vision has limitations — particularly in applications where latency, connectivity, or data sovereignty are critical:

• In autonomous vehicles, milliseconds matter — SaaS can’t compete with on-device inference.
• In remote environments (e.g., mining, agriculture), network access is inconsistent or unavailable.
• In high-security domains (e.g., defense, healthcare), sending data to the cloud may violate regulations.

In these cases, embedded vision or hybrid edge-cloud architectures remain essential.

The Right Tool for the Right Job

SaaS computer vision is not a replacement for embedded AI — it’s a complement. It shines in centralized, connected, and analytic-driven contexts, where the ability to iterate fast and scale instantly delivers real business value.

As the ecosystem matures, expect to see growing adoption in fields like:

• E-commerce: product search and tagging
• Retail: shopper behavior analysis
• Insurance: automated claims from photo uploads
• Healthcare: cloud-based diagnostics and triage support

Why Embedded Vision Matters

Traditional AI relies heavily on cloud infrastructure — powerful, yes, but also vulnerable to latency, bandwidth constraints, and security risks. In mission-critical environments, this is a limitation that can’t be tolerated.

Embedded vision addresses this by putting intelligence directly on the device. The result? Immediate insights, reduced data transmission, enhanced privacy, and full operational autonomy.

“Imagine detecting a drone threat or an equipment failure in milliseconds — not after a server roundtrip. That’s the edge difference.”

Applications in Focus – Some of the most demanding fields:

• Defense: AI vision systems integrated in anti-drone units and mobile defense platforms for instant threat detection.
• Industrial inspection: Offline-capable anomaly detection for factories, remote rail networks, and energy sites.
• Autonomy: Navigation, obstacle avoidance, and object detection for mobile robots and off-road vehicles.
• Smart infrastructure: Real-time analytics for traffic flow, urban pollution, and smart surveillance with full edge privacy.

Modern defense systems require instantaneous threat identification and response. Embedded vision technologies empower autonomous mobile defense units with real-time obstacle detection and rapid decision-making capabilities.​

In industrial settings, unplanned downtime can be costly. Embedded vision systems facilitate continuous monitoring, early anomaly detection, and predictive maintenance, ensuring uninterrupted operations even in harsh environments.​

Aerospace applications demand compact, efficient, and reliable systems. Embedded AI meets these requirements, providing real-time obstacle detection and navigation capabilities essential for autonomous vehicles and microsatellites.​

Embedded vision technologies are at the forefront of smart city initiatives, enabling real-time traffic analysis, urban surveillance, and efficient public transportation systems. By processing data locally, these systems enhance privacy and reduce latency.​

Precision agriculture benefits immensely from real-time data. Embedded vision systems enable farmers to monitor crop health, detect pests, and optimize irrigation without relying on cloud connectivity, promoting sustainable farming practices.​

In retail, understanding customer behavior is key. Embedded vision solutions analyze foot traffic and shopper interactions in real-time, helping retailers optimize store layouts and improve customer experiences. Similarly, in logistics, automated visual inspections streamline package sorting, reducing errors and accelerating delivery times.

Timely diagnostics can save lives. Embedded vision technology assists in medical imaging analysis, providing immediate insights and supporting healthcare professionals in making informed decisions. In emergency scenarios, real-time visual data aids first responders in efficient and effective interventions.​

In this deep dive, we explore how edge vision transforms key industries and use cases — blending performance, resilience, and real-time insight. Here’s how.

Cloud and Edge: Better Together, Not in Competition

Instead of treating cloud and edge computing as opposing choices, organizations can combine both to create smarter, more efficient AI workflows.

• Synergistic Design: Edge devices can handle real-time detection and filtering, while the cloud manages intensive analytics, model updates, and long-term storage.
• Flexible, Optimized Performance: This hybrid model balances low-latency response at the edge with the scale and power of the cloud — ideal for applications that demand both speed and depth.

By leveraging the unique strengths of each, businesses can build image processing systems that are agile today and adaptable for the future.

Looking Ahead: The Future of Embedded Vision at the Edge

At the core of these applications lies a simple truth: AI is only as good as its ability to act — immediately, and reliably. That’s why we believe intelligence must live at the edge, not onliy in the cloud.

As artificial intelligence continues its migration from centralized data centers to decentralized environments, embedded vision is poised to become the cornerstone of real-time perception in critical systems. What was once a domain of niche aerospace and defense applications is now entering mass adoption across agriculture, logistics, automotive, and infrastructure.

According to MarketsandMarkets, the global market for embedded AI is projected to grow from $8.2 billion in 2023 to over $18.6 billion by 2028, driven by demand for on-device analytics, reduced latency, and operational resilience. This acceleration is further fueled by advancements in edge computing, energy-efficient neural processors, and sensor fusion technologies.

The next frontier? Collaborative swarms of autonomous drones, vision-enabled edge robotics, and self-healing smart grids. Future embedded vision systems will not only detect objects but understand context, perform multi-modal fusion with radar or LiDAR, and learn continuously on the edge. Integration with neuromorphic chips, event-based vision, and federated AI training will push the boundaries of what’s possible — securely, in real time, and without reliance on external networks.

From disaster zones to planetary exploration, embedded vision will be the silent enabler of machines that must think independently, act instantly, and operate autonomously — regardless of bandwidth, cloud access, or terrain.

The coming decade won’t just be about intelligent software. It will be about intelligent systems, on-site, on time, and always aware.

—

About Athis Technologies
Athis Technologies Ltd is a European company building real-time, offline-capable embedded vision solutions for critical systems. We deliver edge AI that empowers autonomy, safety, and operational excellence — without compromise.

Athis technologies : A Unified Vision for a Smarter, Safer World

The future of intelligent monitoring doesn’t belong to a single product or architecture — it belongs to ecosystems that are agile, interoperable, and built for the real world.

With its integrated suite of embedded hardware and cloud-native platforms, Athis Technologies delivers more than just point solutions. Whether it’s Sentinel guarding sensitive perimeters, EdgeVision powering industrial reliability, or FleetIQ monitoring vehicles across vast networks, each module is strengthened by its connection to VisionCloud — the centralized AI dashboard that brings everything together.

At the core lies Synergy — the principle that when devices, data, and decisions are unified, the result is greater than the sum of its parts. It’s this synergy that enables Athis to support real-time performance at the edge, centralized insight in the cloud, and seamless orchestration across both.

In an age where milliseconds, security, and scalability all matter, Athis Technologies is building the kind of unified intelligence the world needs — now and into the future.

Visit Athis Technologies

The post Bringing Intelligence to the Edge: How Embedded and SaaS Vision is Transforming Industries appeared first on AthisNews.

Start Date: December 11-12, 2024
Location: Javits Center, NYC, USA

The post The AI Summit New York 2024 appeared first on AthisNews.

Where is it?

Taking place on 9-10 October in Amsterdam, the event aims to gather a large global ecosystem of enterprise, big tech, start-ups, investors and academics to set the global AI agenda.

The post World Summit AI 2024- Amsterdam appeared first on AthisNews.

Start Date: September 8, 9, & 13, 2024
Location: MGM Grand, Las Vegas, NV

The post Next GenAI Conference 2024-Las Vega,NV appeared first on AthisNews.

The partnership leverages Endava’s engineering expertise and Finexos’s AI capabilities to support financial institutions in providing responsible lending.

The Finexos software enriches data analysis, speeds up decision-making, and reduces costs, all while ensuring data security. This alliance aligns with Endava’s commitment to financial inclusion and innovation in the banking sector.

Legacy metrics exclude many borrowers from accessing affordable financial services and expose lenders to unnecessarily high risk as the trend for defaulting borrowers increases. Finexos software, together with Endava’s industry knowledge, aims to increase accessibility to a broader range of customers while providing a much more accurate assessment of borrower risk, resulting in better outcomes for lenders.

“The addition of Finexos strengthens Endava’s banking ecosystem with new and existing AI-based software and also brings to our clients new tools that fuel growth and react to new and future regulatory requirements,” says Lewis Brown, Endava’s head of alliances and partnerships.

While, Matthew Williamson, SVP and industry principal at Endava, added: “This relationship authentically aligns with our core values, using technology and democratising access to financial instruments. Giving individuals and communities the ability to access fair, consistent next-gen real-time credit scoring. Empowering economic growth and stimulus.”

According to Darren Smith, executive director of Finexos, Endava’s decision to plump for Finexos’ advanced AI data science SaaS platform for its customers will “enable it to provide more inclusive, Consumer-Duty-compliant lending decisions while also reducing risk in their lending portfolios”.

The post New Endava and Finexos Partnership Brings Advanced AI to Credit Decisioning appeared first on AthisNews.

A thought leadership piece from the CEO of Kao Data, which develops and operates high performance data centers for advanced computing.

Today Artificial intelligence (AI) has disrupted many of the traditional manufacturing industries, with some of the most rapid changes taking place within automotive, where GPU-accelerated technologies have replaced century-old working practices.

Right now, every aspect of the business is being transformed, from logistics and supply chain management, to design and testing, and even sales and after-service. As autonomous vehicles complete their journey from R&D labs and onto public roads, AI is at the core of the new generation of services and will be pivotal to the industry’s future.

While cars still play the traditional role of getting consumers from A to B, they remain a signifier of lifestyle and status, and are transforming into sophisticated digital products where the software is just as, if not more important than, the hardware. Most of the new AI-based functionality requires high performance computing (HPC) resources for machine learning training, and highly attuned parallel processing that is hard to deliver using traditional CPU-focused architectures.

In fact, the car industry’s demand for AI and machine learning hasn’t been lost on NVIDIA who remains the world’s dominant AI hardware manufacturer.  Since 2019, NVIDIA has launched several innovative solutions aimed specifically at the automotive sector, including its DRIVE Hyperion development platform, DRIVE AGX Orin on-board computer, DRIVE Sim software for virtual testing, and DRIVE Constellation – a hardware platform based on NVIDIA’s DGX servers, designed to run car-related workloads in a high performance data centre environment.

Technologies such as these are driving major advancements within the sector, and British manufacturer Jaguar Land Rover has recently become one of the first luxury car brands to utilize AI-powered compute to transform its product offering. Last month, for example, the company announced it was partnering with NVIDIA to jointly develop autonomous driving and connected services for all future vehicles using NVIDIA DRIVE and will leverage in-house developed data center solutions with NVIDIA DGX for training advanced AI models, with DRIVE Sim software built on NVIDIA Omniverse for real-time accurate simulation.

Design and test

Historically, however, the car design process relied on human expertise, intuition, and iteration, but this approach is fast-being replaced by software and mathematical models.

TÜV SÜD, the German corporation known worldwide for its vehicle crash test certification, offers a wide range of virtual testing services for autonomous vehicles as a means to discover any weak points in the design or its components under real-world conditions, prior to any prototype being produced. Later, AI is also used on the production line, to detect defects and signs of botched assembly, and even track the amount of dust on the car body that could interfere with key consumer finishing aspects like the vehicle’s array of external sensors.

Machine learning systems are also proving invaluable for electric vehicle (EV) research and development, where it is used to improve power efficiency by analyzing the impact of equipment such as heated seats, air conditioning, entertainment systems, and more. This approach enables EVs to drive for longer on a single charge and helps to address one of the key challenges of EV adoption, compared to fossil-fuel powered vehicles.

Living in a simulation

Autonomous vehicles and self-driving cars are not, however, a new concept. The first experiments with algorithm-controlled vehicles took place back in the 1980s, but it took another 40 years to refine sensors and computer vision to a point where they could guide cars along any road, and not just carefully mapped, empty circuits.

Figure 3:Source: Wayve

For many, the initial years of commercial autonomous car development presented a ‘chicken or egg’ scenario where the machine learning models used to control these systems had to be made safe through extensive testing, yet manufacturers were unable to test them on public roads until they were deemed sufficiently safe. 

The answer, as is often the case, came from an unusual and innovative direction – why not create expansive virtual worlds that would replicate all the key characteristics of real roads? This approach allows autonomous vehicles to drive millions of virtual miles, without any danger to drivers or pedestrians. In fact, countless self-driving car technology developers have embraced this kind of virtual testing as part of their R&D, including Alphabet-owned Waymo, Tesla and UK AI Start-up, Wayve.

Wayve has fast become a key name in the autonomous vehicle race, choosing to utilize deep learning technologies and cameras rather than opt for Light Detection and Ranging (LiDAR) technology to guide its vehicles. LiDAR, for example, is a sensor-based technology that creates a map of the environment around it, and was commonplace in the early self-driving vehicle prototypes – being easily identifiable via rotating disks placed on the outside of the car.

To cater for other virtual automotive testing applications, NVIDIA, which is already working with Kao Data on its Cambridge-1 supercomputer, has created a tool called NVIDIA DRIVE Sim being used by Jaguar Land Rover. Introduced in 2021 as part of its Omniverse digital twin platform, the software is intended for autonomous vehicle testing in a wide range of scenarios, and offers groundbreaking render qualities that could, almost, be mistaken for real-life video footage.

Figure 1:Image credit: Kao Data

AI changing the sector – inside and out

Today AI is changing not just how we drive, but how we interact with the car and how it, in turn, interacts with us. A simple example is virtual digital assistants (VDIs). Once only available on smartphones, VDIs have now become commonplace within our homes, and in some car models. Eventually, voice control in the automotive sector will become something customers expect as standard, and that all manufacturers will need to accommodate.

AI is even changing the ownership experience of supercars as more high performance marques launch EV ranges. Part of the familiarization process for new owners is accepting the concept of ‘let the car decide,’ in particular, managing range and power. Through a continuous learning process, the car’s AI system builds a profile of the user’s driving habits and the rate of power consumption.

Further, when selecting a new destination via the navigation system, the vehicle maps out a route which, given the users predicted driving style and consumption, will incorporate the necessary charging points, state of charge on arrival and estimate how long they will wait for charging. Some EVs also incorporate a sophisticated management system with ultra-fast charging capabilities. This feature allows the car to pre-condition and prepare the battery to charge at the fastest rate possible.

Another exciting application of AI is in-car sensing, whereby manufacturers will equip the car with driver and passenger monitoring. This can be used to automate aspects of the in-car experience such as lights or music, but it can also improve safety features. For example, new systems are being developed to detect signs of drowsiness, so if the driver keeps his eyes closed, the car will quickly realise they are likely asleep or in need of medical attention, slow down the vehicle, and park it on the side of the road.

Eventually, cars could also leverage existing infotainment and navigation systems to even recommend a place to stop and have a coffee. Or an alert could be raised if the system detects a child seat left in the car, potentially with a child still in it, or notify the owner that an animal has been forgotten and remains within the vehicle.

What’s clear is that as car technologies evolve, AI will become crucial to both manufacturing and the consumer experience. As autonomous vehicles, for example, become a reality, the ability to store, process and analyze data within a secure and localized, high performance data center will be critical for safety and innovation, and to give automotive manufacturers a competitive industry edge.

The post Artificial Intelligence – A new ally for the automotive industry appeared first on AthisNews.

Visa has launched a new generative AI (Gen AI) powered fraud solution, Visa Account Attack Intelligence (VAAI) Score, to help identify the likelihood of enumeration attacks in card-not-present transactions. 

The latest product launch from Visa comes as malicious actors continue to adopt and leverage more sophisticated technologies – such as botnets and automated scripts – to amplify card testing attacks and exploit transactions at an unprecedented scale and pace.

Now, Visa hopes to sit at the forefront of combatting this rising threat. 

Visa: Protecting against enumeration attacks

Per Visa’s insights, enumeration attacks inflict operational expenses and annual fraud losses of up to US$1.1bn – a significant portion of global fraud. 

To combat this, Visa has updated its VAAI offering with the addition of VAAI Score, a new tool that uses Gen AI components to identify and score enumeration attacks. 

Made available to US issuers first, the VAAI Score will help reduce fraud and operational losses by assigning card-not-present transactions with a real-time risk score – helping detect and prevent enumeration attacks. 

The payments processing giant hopes the effectiveness of this latest innovation is significant, with Visa insights revealing that 33% of enumerated accounts experienced fraud within five days of a fraudster gaining access to payments information. 

Leveraging Gen AI to learn normal and abnormal transaction patterns, Visa’s VAAI Score identifies the likelihood of complex enumeration attacks without compromising the integrity of Visa’s performance and accuracy. 

The launch of VAAI Score comes as Visa hopes to make its services more easily accessible for many by joining the AWS Partner Network, streamlining access to its services for cloud-native fintech, financial institution and enterprise clients. It also follows the release of its new Subscription Manager Service. 

Paul Fabara, Chief Risk and Client Services Officer at Visa, says: “Enumeration can have lasting impacts on our clients and there’s an immediate need for tools that can better detect and prevent these attacks in real-time.

“With the VAAI Score, our clients now have access to real-time risk scoring that can help detect the likelihood of an enumeration attack so issuers can make more informed decisions on when to block a transaction.”

Visa says its new tool has helped reduce the false positive rate by 85% compared to other risk models. VAAI SCORE CAN HELP ISSUERS WITH:

• . Reduced fraud and operational losses
• . Improved cardholder experience
• . Real-time transaction scoring

The VAAI Score model has been trained on over 15 billion VisaNet transactions and has six times the number of features compared to previous VAAI models, helping it better assess suspicious enumeration patterns. 

Using noisy data to train its real-time AI model, Visa’s new risk scoring method derives a two-digit risk score that predicts the likelihood of enumeration to better determine when to approve and decline a transaction. 

Michael Jabbara, SVP and Global Head of Fraud Services at Visa, says: “With access to advanced technology, fraudsters are monetising stolen credentials faster than ever before.

“Enumerated transactions impact the entire ecosystem, and with the VAAI Score, we’re giving our clients a sophisticated tool that can help prevent cardholder accounts from being compromised and stop fraudulent transactions before they happen.”

The post Visa Launches Gen AI-Powered Fraud Solution appeared first on AthisNews.

Start Date: June 6-7, 2024
Location: Phoenix, AZ

The post Machine Learning Week -Phoenix ,AZ appeared first on AthisNews.

The post MLCon 2024 Munich – The Conference for Machine Learning Innovation appeared first on AthisNews.

The summit is for AI practitioners, Hardware, Systems, and Software Engineers and Architects, Data Scientists and AI researchers to unite and discuss the latest challenges in AI infrastructure.

Why attend?

AI Hardware & Edge AI Summit Europe will feature a range of global hyperscaler experts from Google, Microsoft, Meta, and AWS alongside European tech and business leaders who are deploying AI and ML in the sectors where Europe shines, from manufacturing and automotive to finance and retail. 

The agenda has a significant focus on edge AI and edge infrastructure, as well as scaling systems for Generative AI. We’re looking to establish the premier AI infrastructure conference in Europe and unite the world’s best in systems and infrastructure with European application owners and specialists.

We’ll also be featuring Europe’s most innovative hardware, systems and ML startups and showcasing the talent in the region. 

Our 2024 agenda is focusing on making AI faster, more efficient, and more affordable, with sessions covering the latest developments in:

• Server/Cloud AI Infrastructure
• Industrial Edge AI Infrastructure
• Hardware & Systems
• Software Infrastructure & ML Applications

When is

AI Hardware & Edge AI Summit takes place on  18-19 JUNE, 2024 -London

The post AI Hardware & Edge AI Summit 2024 -LONDON appeared first on AthisNews.