FX Vision
Integrate AI powered image and video inspection into any aspect of your business within weeks…without breaking the budget.
What we do
Our FX Vision Solution enables companies to rapidly integrate computer vision into business, quality assurance, or manufacturing processes. We deploy an AI model that is trained to recognize specific items that are unique to your business. Whether it is counting, classifying, identifying errors/faults, or grading, the agent is private and secure, ensuring that no trade secrets leave your company. Once deployed, we offer a range of access (UI or API) and processing options (realtime, batch, or asynchronous) to the computer vision model.
We’ll help you:
Learn the different computer vision options available.
Identify which process would benefit from computer vision.
Draft the specifications and standard operating procedures for image or video capture and storage.
Collect and label a baseline image set to teach your model what to look for in your business.
Determine the best way for your users to interact with the model on a daily basis.
Implement a secure computer vision model only accessible to your company.
Train and enable your teams on how to interact with the model.
The FractionX Solution:
How We Do It
Engage us for a 3 - 5 week process of discovery, collaboration and implementation with our team.
Using a combination of Operational and Technical AI expertise, we will:
Identify, stack rank and document the business requirements driving the need for computer vision.
Draft new computer vision enabled process workflows.
Create and manage a risk register for technology, governance and employee adoption concerns.
Architect and install a cloud based computer vision model on AWS with Amazon Rekognition.
Design a centralized image repository (Dropbox, Google Drive, 0365) for media and setup synchronization to Amazon S3 using AWS AppFlow.
Assist in the collection and labeling of training images with AWS SageMaker GroundTruth.
Fine-tune, test, and deploy a custom AWS Rekognition model based on your company’s requirements.
Integrate the model either into a standalone UI or API your other SaaS systems.
Assist with employee training, communications, and enablement on adoption and integration of your computer vision model.
FX Vision FAQ
-
By leveraging best of breed data tools and AI models from AWS, we can assemble a solution tailored to your requirements at around 80% less time and cost relative to custom “build” AI solutions.
-
First is understanding the limits of computer vision. Most computer vision models can be rolled out quickly for: counting, classifying, and identifying single elements within image or video media. Typically a model can be trained on 100 examples of these images in a few hours.
Complex operations like identifying multiple related elements with dependencies, measuring, or performing qualitative analysis require much more investment in systems design. These models may require hundreds to thousands of images and complex logic with extensive explanations (labels) for the training images.
Second is understanding that computer vision is only as good as the quality of the sample images used to train a model and the quality of the images that are sent for inspection. The model will not perform well on the latter if images veer too much from training images.
-
We host a private and secure version of the model on AWS using the AWS Rekognition service.
-
FX Vision is a services offering we provide to our customers. We deploy FX Vision (Rekognition and GroundTruth) in your AWS infrastructure. If you do not have AWS, we will configure your own instance in your company’s name and give you full control. If you do not have AWS resources on staff, we can provide a monthly service to manage it on your behalf.
-
We use AWS AppFlow and S3 for synchronization of your images. AppFlow can connect securely to any SaaS API to extract your data.
-
We extract data to AWS Simple Storage Service (S3), one of the most secure and well known cloud storage solutions. Your computer vision model will look for and process new images or video in S3.
-
We can provide multiple integration options from a simple UI to upload and inspect images/video, to creating an API your applications can poll, to pushing to another SaaS API endpoint of your choice.
-
We charge an implementation fee to design and build FX Vision. You will then have a monthly fee for AWS cloud usage. For most mid-market companies, this equates to $500 - $2,000 per month depending how often you want to run the model and how quickly it needs to respond (real time, async, or batch).
Why choose FractionX?
There are many (many) companies trying to sell you AI solutions. So why would you choose us?
Experience — 20 years of industry experience are tablestakes to work for FractionX
Certifications – We’re all credentialed with AI certifications (technical or business or both) from an accredited international body.
Culture + Technology – We learned early that Human Centered Design (HCD) is the only path to successful AI technology adoption.
Start + End – We don’t want to bill you forever, our results should manage ourselves out of a job.
Fast and Focused – We don’t waste your time (or ours), let’s collaborate with intention and get to a valuable outcome.
(Yes we do AI, but this ENTIRE page was written the old fashioned way.)
Take the next step
with FractionX.
Learn how FractionX can help solve your mst difficult AI challenges.