Pedab AI Image solution is an offering developed and tailored to solving specific business problems and improve efficiency targeting Manufacturing and Production Industries. It is an easy-to-deploy and easy-to-maintain service that can be accessed for production environments within days. With a user based pricing. Pedab offers a solution that contains one or many AI services, that is consumed in a flexible usage model. With no need for big project commitments.
What is the value?
Pedab AI Image solution can be applied to any use cases where there is, or could be, access to images (or video frames). Images that humans can interpret and label. Most humans typically and statistically can process and identify faults in images to 95% and above – if they are domain experts and the volume of images are low. Our AI solution can do it 24/7 – faster, more precise and at scale.
How does it work?
A typical starting point for Pedab AI Image solution would be defining a use case, selecting a set of representative images to individually classify or label. Pedab’s AI team then develop the algorithm based on the provided and labelled data set.
Validating and testing as the next step is done by using an intuitive framework in the solution, a different set of images is used to verify the accuracy of the algorithm. Validation is always done by our customers. When algorithm is deemed satisfactory, the services can be deployed – on prem or off prem. Choice of deployment options and corresponding infrastructure usually depends on response time requirements. Once the service is in production the algorithm continuously learn from new production data by validation. This enables the solution to infinitely improve.
The offering then covers the full process of AI – Label, Train, Deploy and Learning. All without the need to invest in data science skills. The solution is built for subject matter experts, not data scientists.
Why work with AI?
AI (Artificial intelligence) is a broad term and can have many definitions. The general perception of AI is as problem identification, problem solving and decision making by computers using machine learning techniques such as deep learning models. But one needs to remember that all of today’s applications of AI are niched and boxed to a specific use case – Generic AI is still far away.
Jakob Lagander, Business Unit Manager, +46 766 201 423, firstname.lastname@example.org