View allAll Photos Tagged segmentation

A unique cloud based platform called EOS (Earth Observation System) Crop Monitoring has been announced by the EOS, a Menlo Park based startup in CA, for extracting valuable data such as crop types classification maps, crop yield forecasts, field boundaries, vegetation indices, crop conditions & soil moisture. EOS developed a multi-level deep learning architecture that targets land cover and classifies crop types from multi-temporal, multi-source satellite imagery. The key element of the architecture is an unsupervised neural network that is used for optical imagery segmentation and missing data restoration. This provides reliable algorithms for accurate agriculture statistics estimation and crop yield performance.

Image © Earth Observation System ( EOS )

SATPALDA is an ISO 9001:2008 company and a reseller of multiple satellite data products and allied services. SATPALDA has always tried to push latest Remote Sensing techniques in support for current global issues such as climate change, environment and security and has delivered project-critical geospatial products, including satellite imagery and 3D Modelling services . For more information login: www.satpalda.com

A generative typography using line segmentation, minim length and random angle phasor.

Un Segmentation Fault al estilo Cabañas.

Creating an Effective Marketing Plan with an Opportunity Inventory, Market Segmentation & Competitive Advantage

Better data is becoming available on economic activity within countries, often much exceeding the available information of economic activity across countries. This creates new research opportunities for testing economic theory, analyzing market structures and the sources of market segmentation, and making predictions of how economic shocks propagate across space. The aim of this conference is to bring together researchers from urban economics, industrial organization, health economics, and international trade to study production and trade within and across countries.

Based in Lehfed, a town rich in apple production, Roula established a small enterprise ‘Apple Natura’ to produce apple puree, after initially producing apple jam. As a result of the crisis, challenges in the import of goods had impacted the import of baby food, making the latter a high-demand product. Roula started to produce food for babies and developed a unique recipe for apple and pear puree.

As a result of her participation in WEP, Roula received technical training to help her develop her business skills. The sessions include training on value

proposition, legal aspects, customer segmentation and development, social media, marketing, and sales. The coaching sessions that followed resulted in her rebranding her business to Natura Spoon with the objective of introducing new products and not being limited to apple-based products, and she registered her business at the Lebanese Chamber of Trade & Commerce.

As a result of the WEP in-kind financial support, Roula received a variety of equipment that helped increase production and produce new recipes and products in the baby food industry.

Since she joined the project, her business has been growing exponentially. She has recruited

12 women employees including two nutritionists, and her products are now available in more than 70 pharmacies across Lebanon.

 

Forward Comments and Acronyms explanation added-by BBroomhall,

China Just Pulled Its Own Manhattan Project. Will they still need chips from Taiwan? probably for dominance purposes...

 

An EUV machine (Extreme Ultraviolet lithography machine) is a highly advanced, multi-million dollar system used in semiconductor manufacturing to "print" incredibly tiny, intricate patterns of transistors onto silicon wafers, enabling the creation of today's powerful microchips (like those for AI, smartphones, and high-performance computing)chamber.

Process: The EUV light reflects off the mask, transferring the circuit pattern onto the wafer, repeating the process to build up layers.

Significance,

Advanced Chips: Essential for producing the most advanced logic and memory chips (e.g., 5nm, 3nm, and below) that power modern technology.

Economic & Geopolitical Importance: Command of EUV technology is central to achieving semiconductor self-sufficiency and influences global tech power, with China recently developing a prototype.

Engineering Marvel: These machines are massive, complex, and expensive, requiring a vast ecosystem of suppliers and systems engineering to operate reliably.

This $200M Machine Prints Microchips:

 

en.wikipedia.org/wiki/ASML_Holding#:~:text=It%20is%20the%....

 

ARM: look at WIKI

en.wikipedia.org/wiki/ARM_architecture_family

 

Below Written by Markus Sandelin 2026

  

Or: The West banned the machines. China hired the machinists. Sometimes plans just do not go how you planned them. Ironically, I had been writing this article for a month now, and all research pointed at China being way too far behind. Well…lets check

 

December 2025. Reuters reveals that China completed an operational EUV lithography prototype in a high-security Shenzhen facility. Not through reverse engineering captured ASML machines. Not through some breakthrough in domestic optics manufacturing. Through something far simpler.

 

They recruited the humans who knew how to build them.

 

Former ASML scientist Lin Nan, who headed EUV light source technology in the Netherlands from 2015–2021, filed eight EUV patents in eighteen months — under a new identity, at the Shanghai Institute of Optics. Signing bonuses ran $420,000-$700,000. Housing subsidies. False identity cards. National security classification.

 

The machine fills an entire factory floor. It works. It generates EUV light at 3.42% conversion efficiency. It hasn’t produced functional chips yet — that’s targeted for 2028–2030. But it exists.

 

Every expert consensus said 2035–2040. They were looking at hardware when they should have been watching airports.

 

The Structural Error

Western semiconductor security strategy rested on two pillars. The “silicon shield” thesis: Taiwan’s dominance in advanced chip manufacturing makes invasion counterproductive. And export controls: block ASML from selling EUV systems to China, problem contained.

 

Both pillars shared a hidden assumption. That knowledge lives in machines, not people.

 

ASML’s EUV systems contain over 100,000 components. The mirrors must be accurate to 50 picometers — roughly the width of an atom. The plasma source fires 50,000 tin droplets per second, each vaporized by a 30-kilowatt laser. This complexity was supposed to be the moat.

 

But every specification exists because humans figured it out. The machine is crystallized knowledge. The humans are the source.

 

The same pattern appears in every domain. Organizations obsess over document security while expertise walks out the door every evening. They encrypt files while the knowledge that created them takes a job elsewhere. The artifact isn’t the asset. The artifact is a receipt.

 

China ran a talent acquisition operation in plain sight. Not subtle. Just effective. Because export controls are delay mechanisms. They buy time. They don’t buy safety.

 

The West bought time without using it.

 

What Changed

Before December 2025, consensus estimates placed Chinese EUV capability at 2035–2040. Now a working prototype exists. The timeline compressed by roughly a decade.

 

The prototype remains 2–3 generations behind ASML’s current production systems. The conversion efficiency gap means lower throughput. The machine fills a factory floor where ASML’s fits a room. China still can’t replicate Zeiss’s multilayer mirrors.

 

None of that changes the strategic calculus.

 

The question shifted from “Can China ever catch up?” to “How fast?” Defense planning moved from “permanent advantage” to “degrading advantage with a known expiration date.” Taiwan’s value as a denial target remains, but its value as a capture target increased. If domestic Chinese fabs can eventually produce advanced chips, the calculation for what Taiwan is worth — and what risks are acceptable to take it — changes accordingly.

 

The pattern isn’t unique to semiconductors. Processor architecture tells the same story. x86 dominated server computing for thirty years on the strength of Intel’s manufacturing lead, PC volume economics, and Windows software lock-in. ARM was for phones. Serious computing meant x86.

 

Now ARM powers over half of new cloud deployments. The conditions that protected x86 systematically reversed. Manufacturing advantage moved to TSMC, accessible to all ARM licensees. Volume advantage shifted to smartphones. Software lock-in dissolved as cloud-native development reduced Windows dependency.

 

What looked like permanent architectural dominance was a phase. The moat didn’t hold because moats don’t hold. They erode, get bridged, or someone builds a tunnel underneath.

 

The Dependency Structure

The AI compute stack that powers Western technological advantage contains a specific shape of fragility.

 

Every major AI system depends on one GPU architecture (NVIDIA), manufactured at one foundry (TSMC), located in one place (Taiwan), using machines from one supplier (ASML), with precision optics from one manufacturer (Zeiss). Four single points of failure in sequence. Remove any node and the capability degrades or disappears.

 

This isn’t hidden information. It’s the structure everyone knows and nobody fixed.

 

NVIDIA’s position creates particular tension. Dominant market share in AI accelerators. A software ecosystem — CUDA — that four and a half million developers are locked into. Aggressive capacity reservation that secures most of TSMC’s advanced packaging through 2025. The dependency runs deep enough that when Meta reportedly considered Google’s TPUs for future data centers, NVIDIA’s stock dropped four percent on the news alone.

 

The responses emerging suggest the dependency is recognized if not resolved. Google’s TPUs now train Anthropic’s Claude. AWS Trainium claims significant cost advantages. AMD’s ROCm has narrowed performance gaps enough that seven of the top ten AI companies deploy AMD accelerators alongside NVIDIA. None of this eliminates the concentration. It creates hedges against it.

 

OpenAI’s infrastructure commitments reveal the scale of what’s being built on this fragile foundation. Over a trillion dollars in spending through 2035, spread across multiple suppliers but unified by a common dependency on hardware that originates in Taiwan. The Stargate Project alone commits $500 billion to build ten gigawatts of AI data centers. Power is the bottleneck — gas turbines filling gaps the grid can’t supply.

 

This is infrastructure for an AI capability that assumes the current supply chain continues functioning. That assumption now has a visible expiration date.

 

The Coming Fragmentation

Technology competitions resolve in different ways. Sometimes one standard wins completely. Sometimes markets segment until competitors stop fighting over the same territory.

 

The AI compute ecosystem appears headed for segmentation along geopolitical lines.

 

China must develop domestic alternatives regardless of technical merit. The motivation isn’t commercial — it’s strategic necessity. Huawei open-sourced its neural network framework in August 2025. Chinese AI models launched with domestic accelerator support on day one. The ecosystem will exist because it has to, competitive or not.

 

Western hyperscalers face different pressures toward the same fragmentation. Monopoly suppliers eventually extract monopoly rents. Google, Amazon, and Microsoft are all developing custom silicon not because they enjoy manufacturing complexity, but because dependency on a single supplier creates unacceptable leverage.

 

The workload mix is shifting in ways that accelerate this. Training AI models requires the deep optimization that CUDA provides. Inference — running trained models — is more amenable to alternatives. As inference grows from minority to majority of compute workloads, the lock-in weakens.

 

The likely outcome isn’t one winner. It’s regional technology blocs with limited interoperability. Western infrastructure running mixed NVIDIA and alternatives. Chinese infrastructure running domestic stacks. Emerging markets splitting based on cost sensitivity and political alignment.

 

Format wars end not with surrender but with the market drawing borders until both sides stop contesting the same ground. Those borders are being drawn now.

 

What Sovereignty Actually Requires

The export controls bought approximately six years of development time. That’s not nothing. It’s also not victory.

 

Technology restrictions do specific things. They slow adversary development. They create incentives for domestic alternatives. They signal strategic importance — which focuses opposing effort. What they cannot do is create permanent barriers. That has never happened. In any domain. Historically.

 

The pattern is consistent. Restrictions buy time. Time gets wasted. The capability gap closes. Surprise ensues.

 

Durable technological sovereignty requires different investments.

 

Velocity of innovation matters more than blocking followers. Staying ahead is sustainable. Preventing catch-up is temporary. ASML didn’t achieve dominance through export controls — it achieved dominance through decades of capability investment that competitors couldn’t match. Controls came after the gap existed, not before.

 

Ecosystem integration creates resilience that individual technologies cannot. ASML depends on over 1,200 specialized suppliers. That network is harder to replicate than any single machine. Complexity becomes a feature when it emerges from genuine interdependence rather than artificial restriction.

 

Human capital retention costs less than blocking departure. The engineers who built China’s prototype weren’t captured or coerced. They were recruited with money, housing, and opportunity. Talent retention is a compensation and culture problem. It has solutions that export controls don’t provide.

 

Geographic distribution reduces single-point risk. TSMC began Arizona production in 2025. Intel is building in Ohio. Samsung is expanding in Texas. None of this is sufficient relative to Taiwan concentration, but all of it represents movement in a direction that should have started a decade earlier.

 

Alternative development prevents lock-in from becoming vulnerability. Open-source frameworks, competing architectures, multiple viable paths to the same capability — these aren’t just market competition. They’re strategic hedges against ecosystem capture.

 

The West had six years of bought time. It could have invested in talent retention, funded open alternatives, built distributed manufacturing capacity, treated the window as finite.

 

Instead: self-congratulation about the controls, consolidation of existing dependencies, continued concentration in Taiwan, and the assumption that the gap would never close.

 

Sovereignty built on preventing others from having knowledge runs on a countdown timer. The timer was always visible. Looking at it was optional.

 

The Pattern Repeating

Ukraine is now the world’s most advanced live-fire laboratory for defense technology.

 

Drone warfare at scale. Electronic warfare that evolves weekly. AI-assisted targeting systems built under fire. Distributed command and control that survives decapitation strikes. Civilian-military tech integration that NATO has theorized about for decades but never implemented. All of it battle-tested under conditions no simulation can replicate.

 

The engineers and operators who built these systems didn’t learn from textbooks. They learned by iterating faster than the enemy could adapt, with lives depending on each software push. That expertise — the hard-won knowledge of what actually works when systems face real jamming, real counter-drone measures, real attempts to kill the operators — exists nowhere else on earth.

 

The West’s support pattern should concern anyone who understood the ASML lesson.

 

Delayed weapons deliveries. Conditional aid packages. Restrictions on how provided systems can be used. Allies calibrating support to “enough not to lose, not enough to win.” A persistent message that Ukrainian victory is desired but not prioritized. Economic exhaustion setting in after three years of war.

 

These are the conditions that make talent extraction attractive.

 

China’s semiconductor recruitment playbook is proven and replicable. Signing bonuses that represent a lifetime of Ukrainian military salary. Housing in cities not under missile attack. The chance to work on systems that will actually get built and deployed, not stuck in procurement limbo. For engineers who’ve spent three years surviving on Western uncertainty, the pitch writes itself.

 

The assumption that Ukrainian defense talent will remain loyal to the West rests on the same error as the assumption that ASML knowledge would stay in the Netherlands. Loyalty follows from earned trust and mutual investment. It does not follow from geography or shared values alone — especially when the values appear negotiable and the investment appears grudging.

 

The knowledge of how to build defense systems that work in modern warfare lives in Ukrainian engineers and operators. Not in the drones themselves. Not in the targeting algorithms. Not in the command systems. In the humans who built them, tested them, died alongside them, and rebuilt them better.

 

If the West continues treating Ukraine as a buffer state to be supported at minimum viable cost rather than an ally whose expertise represents irreplaceable strategic value, the same extraction pattern will repeat. Different industry. Same blindspot. Identical outcome.

 

China has demonstrated it will pay whatever it takes for knowledge that matters. The question is whether the West understands what Ukrainian defense expertise is worth before someone else makes a better offer.

 

The Pattern Beneath

China’s Manhattan Project didn’t emerge from nowhere. It came from patience, resources, and the humans who knew the answers.

 

The same pattern appears everywhere once you look for it. Institutions that secure documents while knowledge leaves every evening. Organizations that patent processes while the engineers who invented them take jobs elsewhere. Nations that restrict hardware exports while talent flows freely across borders.

 

The artifact is never the asset. The artifact is what the asset left behind.

 

What looks like an insurmountable wall is always, eventually, an expensive door. The only question is how motivated the people on the other side are. China’s latest semiconductor fund: $37 billion. Black-budget projects beyond that. Whatever it takes to recruit the right humans.

 

That’s how motivated.

 

The West fortified the castle. China recruited the castle builders.

 

The wall was always a door. The hinges just weren’t visible from the inside.

 

The author has spent way too many years watching institutions mistake structural barriers for permanent advantages. He’s consulted on transformation for NATO, EU institutions, and organizations that genuinely believed complexity would protect them forever. He’s tired of watching intelligent people confuse artifacts with knowledge, and restriction with security. His rubber chicken remains unimpressed.

 

AI

Nvidia

China

Defense

Technology

2.9K

  

60

  

Markus Sandelin

Written by Markus Sandelin

709 followers

·

202 following

Combining the internet with the defense industry

  

Follow

Responses (60) Fwd, To

B Brmhall

B Brmhall

What are your thoughts?

Cancel

Respond

Astoria Bob

Astoria Bob

 

5 days ago

   

Europe Classical Style Furniture

 

European classical furniture, the craftsmen in the 17th century to the 19th century made specifically for the royal family, noble furniture by hand.The style of the furniture after hundreds of years of change, there has been no change it the characteristics of the work, keep improving.Full decoration is the biggest characteristic of European classical style furniture.CDG furniture is a commercial furniture Manufacturers.Mention European classical furniture, always reminiscent of the deep cultural inside information, a few hundred years down the transmission of culture, in the furniture of the perfect embodiment.European classical furniture mainly includes Italian style furniture, French style furniture, German, Spanish style furniture.

 

European classical furniture is mainly divided into the baroque and rococo furniture “, late again more concise “new classic furniture”.

 

The Baroque Style

 

Features: with curved surface, twists and turns, such as flow, interspersed with flexible exaggeration stressed to create special artistic effect, to render the mysterious religious atmosphere and the beauty of floating illusion.

 

The Rococo Style

 

The rococo style with its uneven light delicate curve is famous.Adornment modelling using the curve of the shells, segmentation, buckling and bending shape composition of trival, decorative capabilities, accepteth colorful and varied, and the utilization of the volume grass grain in China, has a light, flowing, extend outward adornment effect.CDG furniture is a metal furniture manufacturers.The rococo furniture color is relatively soft, cream-colored, white soft decorative pattern design is the main color.

 

Italian furniture is occupying the important status in the history of world furniture, the design capacity is famous worldwide.Britain’s Buckingham Palace, the White House can see the figure of Italian furniture, the royal style can be seen;In the international market, the Italian furniture is modern and luxurious pronoun.

 

French style, refers to the French national building and furniture style.Mainly includes the baroque style (style of Louis xiv) French rococo style (style of Louis xv) neoclassical style (Louis xvi style) style of caesarism, etc., is the top of the European furniture and architectural culture.

Better data is becoming available on economic activity within countries, often much exceeding the available information of economic activity across countries. This creates new research opportunities for testing economic theory, analyzing market structures and the sources of market segmentation, and making predictions of how economic shocks propagate across space. The aim of this conference is to bring together researchers from urban economics, industrial organization, health economics, and international trade to study production and trade within and across countries.

I dream of a world where we are all deeply dedicated to nurturing ourselves, eachother, the planet and a world we can leave behind to our children; to restore the natural and intended balance of our universe; to it's beautiful, whole & delicately complex design of precision.

 

Create a United Planet, a global platform that unifies all resources and individuals towards a common purpose - ending suffering and exploitation of our children. The global platform reflects the cultural creatives, the innovative and open minded individuals in our government , education, social & legal systems, our families, communities and corporations.

 

All individuals unified on this global platform operate from the rooted wisdom that there is more than enough of everything, for everyone. Members operate without borders or segmentation of natural or financial resources.

 

Within the body of a United Planet, the spirit of generosity, compassion, empathy, collaboration, inclusiveness, trust, love and wisdom are the core root values and life force behind all decisions made.

 

[OW photo id 14155]

A generative typography using line segmentation, minim length and random angle phasor.

A generative typography using line segmentation, minim length and random angle phasor.

Changing the natural habitat introduces conventions. These are embodied in usual traffic signs and also in functional zones or administrative and community related institutions. And for their better comprehension and for a quicker capacity towards human adaptation to artificial space, there are maps. Maps are objective indicators that help you clear your doubts or missing information.

“Danube Polluted” and “Shop to gray” are maps, the first one of Novi-Sad, dealing with the pollution of the Danube, the second one of Belgrade, dealing with the segmentation of the city, divided in commercial, natural, new or old areas, being given a different color.

A lineament is a linear feature in a landscape which is an expression of an underlying geological structure such as a fault. Accurate geological features mapping is a critical task for oil exploration, groundwater storage and understanding the mechanisms of environmental disasters. Multispectral satellite imageries are very useful in this regard. Image processing tools used for lineament feature detection are image enhancement technique, edge detection and segmentation, spatial domain filtering techniques, band-rationing and lots of filtering processes. Digital Elevation Model (DEM) provide better idea regarding topographic elevation and thus help in delineating geological structures.

 

SATPALDA is an ISO 9001:2015 certified leading provider of geospatial products and services. The company is also a reseller of multiple satellite data products and has a proven track record of delivering project-critical geospatial products, including satellite imagery, UAV imagery, elevation models, LULC maps, planimetry, terrain solutions and many more. For more information login: www.satpalda.com.

From this infographic learn about the journey for partners on blockchain based loyalty program and how it opens up different avenues across touch points, for analytics, segmentation, personalization, sales prediction, cross-selling, up-selling, and so on.

Yes, cultural segmentation is alive and well in the US!

Internet Download Manager is the choice of many, when it comes to increasing download speeds up to 5x. IDM has a clever download logic accelerator that features intelligent dynamic file segmentation and incorporates safe multipart downloading technology to increase the speed of your downloads....

 

psncodegames.com/idm-6-20-full-crack-internetdownloadmana...

Train your machine learning model with our high-quality human segmentation dataset and unlock new levels of accuracy and precision in human detection and tracking.

Our investigators observing the circumstance over the globe clarifies that the market will create gainful possibilities for makers post COVID-19 emergency. The report means to give an extra outline of the most recent situation, monetary stoppage, and COVID-19 effect on the general business.

 

Shower Enclosure and Cubicles Market is growing at a faster pace with substantial growth rates over the last few years and is estimated that the market will grow significantly in the forecasted period i.e. 2019 to 2026

 

Shower Enclosure and Cubicles Market Top Key players:

 

•Huppe

•Jaquar

•Duravit

•Kohler

•LIXIL Group

•Porcelanosa

•Masco

 

Shower Enclosure and Cubicles Market Segmentation Analysis

 

•By Type

•By Market Dynamics

•By Region

 

SD-WAN's cloud performance capabilities are vast and quite feature rich.

They provide features like:-

1. Network Segmentation

2. Better application performance

3. Network Security

4. LAN Management

 

Visit Blue Fox Group to know more about SD-WAN's cloud capabilities.

A generative typography using line segmentation, minim length and random angle phasor.

A generative typography using line segmentation, minim length and random angle phasor.

A generative typography using line segmentation, minim length and random angle phasor.

At vyakar, you will get the best in class services alike data standardization, fuzzy matching software, and lead segmentation - most important is better together.

Bing, the search engine of Microsoft with the help of its mapping team added about 125 million building footprints in USA to open street map community. There team used a Microsoft developed CNTK toolkit for this purpose. They detected building footprints from the Bing imagery. Two primary stages involved in this process were 'Semantic Segmentation' and 'Polygonization'. As the footprints were detected from Bing imageries they consist of noise distortion and suspicious data which were erased from predictions and then a Polygonization algorithm was applied to detect the building edges and angels. As a result of this process proper buildings were created.

Building footprints are one of the most important features of a map. These are used in urban planning, LULC maps, Taxation, Aviation and many more other applications. SATPALDA successfully executed many building footprint based projects and we are capable of producing 3D building footprint too. The company is also a reseller of multiple satellite data products and has a proven track record of delivering project-critical geospatial products, including satellite imagery, UAV imagery, elevation models, LULC maps, planimetry and many more. For more information login: www.satpalda.com.

 

Conventional gray scale segmentation on the artificial image.

Better data is becoming available on economic activity within countries, often much exceeding the available information of economic activity across countries. This creates new research opportunities for testing economic theory, analyzing market structures and the sources of market segmentation, and making predictions of how economic shocks propagate across space. The aim of this conference is to bring together researchers from urban economics, industrial organization, health economics, and international trade to study production and trade within and across countries.

A generative typography using line segmentation, minim length and random angle phasor.

Company user segmentation poster shot

Creating an Effective Marketing Plan with an Opportunity Inventory, Market Segmentation & Competitive Advantage

From the museum label:

Kim Whanki is one of many twentieth-century Korean artists who admired and collected Joseon decorative arts. Especially fond of porcelains, he frequently featured them in his paintings, sketches, and magazine illustrations. Moon and Jar exemplifies Kim's keen sense of color, adroit brushwork, and use of segmentation. While the background may seem abstract, the title guides us to read the circle as the moon, its diameter echoing the jar's. Kim has in fact been credited (anecdotally) with giving this type of vessel the name "moon jar."

The painting treats the jar as an art object, rather than as an auspicious symbol or visual pun as in earlier Korean still-life painting, but it also carries a poignant nostalgia. Much of Kim's collection was lost or destroyed during the Korean War (1950- 53). Later, while in Paris (1956-59) and New York (1963-74), he found comfort and inspiration in Joseon objects, which he continued to collect.

1 2 ••• 36 37 39 41 42 ••• 79 80