Were you unable to attend Transform 2022? Check out all Summit sessions in our on-demand library now! Look here.
Edge devices must be able to process delivered data quickly and in real time. And edge AI applications are only effective and scalable if they can make highly accurate image predictions.
Take the complex and mission-critical task of autonomous driving: All relevant objects in the driving scene must be taken into account – be they pedestrians, lanes, sidewalks, other vehicles or traffic signs and traffic lights.
“For example, an autonomous vehicle driving through a crowded city must maintain high accuracy while operating in real time with very low latency; Otherwise, driver and pedestrian lives may be at risk,” said Yonatan Geifman, CEO and co-founder of deep learning company Deci.
The key to this is semantic segmentation or image segmentation. But there is a dilemma: semantic segmentation models are complex and often slow down their performance.
MetaBeat will bring together thought leaders on October 4th in San Francisco, California to provide guidance on how Metaverse technology will transform the way all industries communicate and do business.
“There’s often a trade-off between the accuracy and the speed and size of these models,” said Geifman, whose company this week released a suite of semantic segmentation models, DeciSeg, to help solve this complex problem.
“This can be a barrier to real-time edge applications,” Geifman said. “Creating accurate and computationally efficient models is a real pain point for deep learning engineers who go to great lengths to achieve both the accuracy and speed to accomplish the task at hand.”
The power of the edge
According to Allied Market Research, the global edge AI (artificial intelligence) market size will reach nearly $39 billion by 2030, representing a compound annual growth rate (CAGR) of nearly 19% over 10 years. Meanwhile, Astute Analytica reports that the global edge AI software market will reach more than $8 billion by 2027, growing at a CAGR of nearly 30% from 2021.
“Edge computing with AI is a powerful combination that has promising applications for both consumers and businesses,” said Geifman.
For end users, this means more speed, improved reliability and a better overall experience, he said. Not to mention better data protection, since the data used for processing stays on the local device – mobile phones, laptops, tablets – and does not have to be uploaded to third-party cloud services. For companies with consumer applications, this means a significant reduction in cloud computing costs, Geifman said.
Another reason why Edge AI is so important: communication bottlenecks. Many machine vision edge devices require high-performance analytics for high-resolution video streams. However, when the communication requirements are too large in relation to the network capacity, some users do not get the analysis they need. “Hence, moving computation to the edge, even partially, will enable large-scale operations,” Geifman said.
No critical compromises
Semantic segmentation is key to edge AI and one of the most commonly used computer vision tasks in many industries: automotive, healthcare, agriculture, media and entertainment, consumer applications, smart cities, and other image-intensive implementations.
Many of these applications “are critical in that it can be a matter of life and death to get a correct real-time segmentation prediction,” Geifman said.
Autonomous vehicles for example; another is cardiac semantic segmentation. For this critical task in MRI analysis, images are divided into multiple anatomically meaningful segments that are used to estimate critical factors such as myocardial mass and wall thickness, Geifman explained.
Of course, there are examples that go beyond mission-critical situations, he said, like virtual background capabilities for video conferencing or smart photography.
Unlike image classification models, which are designed to identify and label an object in a specific image, semantic segmentation models assign a label to each pixel in an image, Geifman explained. They are typically designed using an encoder/decoder architecture structure. The encoder incrementally downsamples the input while increasing the number of feature maps, thereby constructing informative spatial features. The decoder receives these features and converts them step by step into a full resolution segmentation map.
And while it’s often required for many edge AI applications, there are significant barriers to running semantic segmentation models directly on edge devices. These include high latency and the inability to deploy models due to their size.
Not only are highly accurate segmentation models much larger than classification models, Geifman explained, they are also often applied to larger input images, which “increases their computational complexity quadratically”. This results in slower inference performance.
For example, flaw inspection systems running on assembly lines that need to maintain high accuracy to reduce false alarms, but without sacrificing speed, Geifman said.
Lower latency, higher accuracy
The DeciSeg models were automatically generated by Deci’s Automated Neural Architecture Construction (AutoNAC) technology. The Tel Aviv-based company says these “significantly outperform” existing publicly available models, including Apple’s MobileViT and Google’s DeepLab.
As Geifman explained, the AutoNAC engine takes into account a large search space of neural architectures. Searching this space takes into account parameters such as base precision, performance goals, inference hardware, compilers, and quantization. AutoNAC attempts to solve a constrained optimization problem while achieving multiple goals—that is, preserving baseline accuracy with a model that has a specific memory footprint.
The models deliver more than two times lower latency and 3% to 7% higher accuracy, Geifman said. This enables companies to develop new use cases and applications on edge AI devices, reduce inference costs (since AI practitioners no longer have to perform tasks in expensive cloud environments), enter new markets and reduce development times, Geifman said. AI teams can solve deployment challenges while achieving desired accuracy, speed, and model size.
“DeciSeg models enable semantic segmentation tasks that previously could not be performed on edge applications because they were too resource intensive,” said Geifman. The new models “have the potential to transform industries altogether.”
VentureBeat’s mission is intended to be a digital marketplace for technical decision makers to acquire knowledge about transformative enterprise technology and to conduct transactions. Discover our briefings.