Deploying AI Models at the Edge

LeapMind introduces a new AI image processing model, bringing quantization tech to edge devices

[Image Credit: MARTECH]

Tokyo, Japan-based Leapmind, known for its ultra-low-power AI inference accelerator, Efficiera, aims to solve the challenges faced by modern AI systems. Leveraging through its low-bit quantization technology, the IP vendor has announced a new AI image processing model that can efficiently operate on edge devices to improve the image quality of noisy images.

As the market for optimized machine learning models for computer vision applications has grown, researchers have continued to work on improving performance through custom AI models. This is because edge devices face the challenge of dealing with increasing volumes of incoming data and high-computational load to serve real-time applications.

In most cases, to increase the performance and accuracy of the processed images, there is a high-cost component aspect involved. As the input images travel a long design flow, there are high chances of noise added to the image. For AI image processing, the software/hardware ecosystem with low-bit quantization technology added on top of the high performance of Efficiera v2 has made it possible to overcome these challenges and efficiently operate even in a video camera, according to the company.

“As far as we searched, we are the world’s first in bringing this Image processing AI model into the product by low bit quantization technology. This model is a product that can be put into practical use only because of LeapMind focusing on both hardware and software development and we believe that it shows the new value of extremely low bit quantization,” said Hiroyuki Tokunaga, CTO of LeapMind.

Leapmind said its solution is viable for deployment in industrial camera-like security systems requiring edge computing of the incoming video/image processing to improve accuracy and object recognition. Employing such models in edge devices can immensely improve the way today’s smart camera projects work while reducing the computational cost.

According to the press release, the model is expected to be available for evaluation from February 2022.

The value of AI image processing models that not only enhance image quality but also reduce the hardware/software set-up cost involved means that there will be significant investment activity for the foreseeable future. So far in 2022, the edge AI and computer vision market has seen companies like Nota announcing that it closed a $14.7 million Series B funding round to optimize AI models. With this funding news, it was evident that the industry and the customers are looking for better AI models for image processing to aid the huge market for computer vision applications.

Article Topics

 |   |   |   |   | 

Comments

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Featured Edge Computing Company

Edge Ecosystem Videos

Deploying AI Models at the Edge

“Deploying

Latest News