Easy-to-use development platform brings together pre-trained computer vision models, innovative APIs, starter applications and edge environments
New product features include pose estimation, semantic segmentation and support for NVIDIA Jetson Nano
All computer vision developers are encouraged to create a free account now at alwaysai.co
SAN DIEGO, Calif. (February 12, 2020) – alwaysAI, a software company dedicated to making computer vision (CV) accessible to all developers with its innovative and easy-to-use platform, announced today that its beta program is open for all software developers to quickly create and deploy CV applications on the edge.
The move builds on the momentum of its recent private beta program, which attracted more than 1,600 developers since it opened in October 2019.
“We see enormous interest in computer vision and how it can drive new value for enterprise applications and developers – it makes AI and IoT come alive in the real world,” said Marty Beard, co-founder and CEO of alwaysAI. “We love evangelizing the value of this exciting new technology, and I believe we are fundamentally changing the way developers create, prototype and deploy computer vision applications at the edge.”
Developers, from beginners to experts alike, are encouraged to open their free account at alwaysai.co and start creating, prototyping and deploying CV apps on ARM-based devices including cameras, drones, wearables, industrial monitoring equipment and transportation units.
Making Development Easy
In the open beta, developers get immediate access to a growing, searchable catalog of pre-trained computer vision models, a full set of CV primitives including image classification, object detection, tracking, counting, face detection – and now human pose estimation and semantic segmentation, as well as an expanding array of supported edge environments.
In addition, developers get an open channel of communication with alwaysAI’s rapidly growing developer community and direct access to the engineering team.
alwaysAI’s accessible and user-friendly platform enables developers to create and deploy computer vision applications in three simple steps:
- Step 1: Choose a computer vision model. Browse alwaysAI’s growing catalog of high-quality models, which are pre-trained, benchmarked and guaranteed to work across a wide variety of edge environments.
- Step 2: Develop an application. Use alwaysAI’s robust, model and hardware-independent APIs to rapidly prototype and refine applications.
- Step 3: Deploy an application to the edge. Use alwaysAI’s command line interface and performance assessment tools to test and deploy their applications on ARM32, ARM64 and x86 platforms, and with popular accelerators.
And with alwaysAI, inferencing happens on the edge, so there is lower latency and no required cloud hosting or inference charges – a significant cost and time savings.
Growing Demand from Developers
Co-founders Marty Beard and Steve Griset started alwaysAI with the intention to democratize computer vision and help all enterprise developers leverage CV in practical and affordable ways.
“We have seen communications – like messaging and speech – get automated and proliferated across a wide variety of end-points. But vision – arguably the most powerful human attribute – has simply been too difficult for technologists to implement and deploy,” Beard said. “With the open beta program, we are broadening access and offering new features that make it even easier and more powerful for the everyday developer.”
Developers from a wide variety of backgrounds and industries agree:
- “We have found the alwaysAI platform to be a powerful tool in creating computer vision applications,” said Dave Tunney, managing director at SparroWatch. “alwaysAI is helping us solve real-world problems with computer vision. The alwaysAI platform is simple to use, and we are happy with its performance. The guys at alwaysAI have been great in supporting our development.” – AI systems integrator SparroWatch Ltd.
- “At Seafloor Systems, our challenges are uniquely positioned to use a computer vision platform like alwaysAI because the very nature of our work is on the edge,” said Marcos Barrera, lead robotics and AI research engineer at Seafloor Systems, Inc. “We not only have to contend with the physical limitations of working on the water but also have to balance between the constraints of power and data acquisition on remote unmanned autonomous surface vessels. alwaysAI gives us the freedom to be innovative in our solutions because their team built the platform with flexibility at the core.” – Maritime drone manufacturer Seafloor Systems Inc.
- “Our platform works seamlessly with alwaysAI so developers can deploy, manage and update their computer vision on fleets of connected IoT devices,” said Alison Davis Riddell, chief operating officer at balena. “We're excited to be working with alwaysAI and their powerful developer platform, which makes it fast and easy to create computer vision applications at the edge.” – Software company balena Inc.
- “From IoT to robotics and automation, alwaysAI continues to be our go-to platform of choice for computer vision development,” said Dean Pfutzenreuter, co-founder and chief technology officer at productOps. “It's quick, easy to use, stable and highly flexible, even for the most challenging computer- vision projects on the edge. Using alwaysAI helps us get our computer vision work for clients done faster than ever before.” – Software development consultancy productOps Inc.
- “alwaysAI allowed my team and I to focus on what mattered,” said Christian Piper, Vision Systems’ FIRST robotics team lead. “The platform provided a way to implement complicated features easily. Using the vision system, even with little experience on how machine learning and image analysis works, we were able to implement a vision system that was more robust than any other team in our worldwide competition.” – Not-for-profit STEM engagement program FIRST (For Inspiration and Recognition of Science and Technology).
- “We are currently building an in-home ambient monitoring sensor and app that tracks a person's movements and analyzes them given specific metrics for providing preventative and health monitoring solutions,” said Raffy Rabin, chief solutions architect at Cherish Health. “The alwaysAI platform and system is very easy to use, and the development team is very easy to work with. I was able to prototype and get a working demo up and running in less than a day.” – AI health solutions company CherishHealth.
New Computer Vision Features
alwaysAI’s computer vision software is now available on NVIDIA’s Jetson systems. The combination of alwaysAI's software and NVIDIA's Jetson hardware will provide intelligent sight to devices that run autonomous machines, smart cities, retail services and other advanced computer vision applications.
NVIDIA’s Jetson Nano is a small, powerful computer that lets a developer run multiple neural networks in parallel for applications like image classification, object detection, segmentation and speech processing. The Jetson Nano is the ideal platform for creating high-performance deep learning, computer vision projects at the edge.
The alwaysAI platform also makes it easy to build, test and deploy computer vision applications for autonomous driving applications, including a pedestrian and bicyclist detector equipped with semantic segmentation. Autonomous vehicles need to determine how far away pedestrians and bicyclists are, as well as their intentions.
With semantic segmentation, detections are done pixel-by-pixel, rather than with bounding boxes. In certain scenarios – like foot and bicycle traffic in bustling urban areas – the autonomous vehicle needs much more detailed information about the exact location of a pedestrian or a bicyclist. alwaysAI makes that fast and easy.
In opening up the beta, alwaysAI carefully listened to its growing developer base, offering more of the powerfully optimized tools they want. The alwaysAI platform recently released an easy-to-deploy image for both Raspberry Pi 3B+ and the Raspberry Pi 4. Enterprise software developers and hobbyists alike are tapping into the alwaysAI platform to get their edge computer vision projects up and running.
“alwaysAI is an awesome product ... it makes computer vision development on the edge simple,” said Tomas Migone, hardware hacker in residence at balena. “The tools are easy to use, and the documentation is straightforward. Developing with alwaysAI is a great experience. I'm looking forward to continuing using it for computer vision projects.”
For more information about alwaysAI, or to join the company’s open beta program, visit alwaysai.co.
alwaysAI (alwaysai.co) brings deep learning computer vision to embedded and IoT devices. By providing a professional catalog of pre-trained models, innovative set of computer vision APIs, and growing array of supported edge environments, alwaysAI accelerates the time it takes to get a computer vision app up and running. Based in San Diego, alwaysAI is led by a team of technology veterans who are passionate about democratizing access to computer vision. Co-founders Marty Beard and Steve Griset have more than 40 years of combined experience in enterprise software, mobility, cloud applications and cybersecurity.