Image recognition is part of a bigger idea or workflow
We can train a model to identify and extract specific information from images like document types, logos, objects, products or anomalies.
If your images cover a specific subject only, we can train a visual search model, optimized for your particular image domain.
You have the requirements, we have the tools.
Together with you, we clearly understand and describe the problem to be solved.
We analyze your available data and discover and prepare the structure needed for training.
We train a first model to retrieve a baseline accuracy from which to improve, if necessary.
We take measure to improve the results to fit your goal. This includes data cleansing, data curation, model configuration and hyper parameter tuning.
We package and deliver your production ready AI model as API module which can be easily integrated in your existing systems and workflows.
Your AI models run in your infrastructure
You custom AI model is wrapped with an API and packaged as Docker image.
We provide your custom AI model as module for our Apache Solr plug-in.