Image Recognition

Image and Video Recognition

Computers can be taught to accurately identify what’s in pictures, and even videos, faster than ever — but they need massive amounts of data to do it. A conventional neural network attempting to process even a relatively small image (such as 30×30 pixels) would require 900 inputs and nearly half a million parameters. While that might be manageable for a reasonably powerful machine, once the images become larger (say 500×500 pixels), the number of inputs, parameters, and overall compute power required increases to absurd levels. Even more so, companies are faced with trying to scale their operations for image recognition on cloud services such as AWS or GCP only to realize that prolonged usage of these services for image classification can be prohibitively expensive.

Cirrascale Cloud Services has been using GPUs for image recognition, such as the NVIDIA® Tesla® V100 32GB GPU Accelerator, to make groundbreaking improvements across image classification and processing applications. Since so much overall GPU compute power is needed, the Cirrascale Cloud Services platform is a much better choice for those companies needing cloud-based, dedicated servers. You'll get charged one flat rate per month and that's it. No ingress or egress fees, no overage charges, no surprises... and we provide a level of support that goes beyond other providers.