Braskem
Smart Gels: AI classification of gels in polymer films
Braskem built AI models on Dataiku to classify gels in polymer films, with an estimated 30% cut in lab analysis time
Problem
Gels are defects in polymer films, and classifying them in the lab takes analyst time.
Approach
Braskem's team built image-processing scripts, a gel detection model and gel classification models in one Dataiku project. Users review the model classifications and work through an application, and the results are stored. At the time of the post the project was in progress after about six months of development.
Results
Takeaway
The savings are the team's own estimates made during development, not measured results.
Categories
- Size
- Large enterprise
- Industry
- Chemicals and materials
- Business function
- Production and qualityIT and R&D
- Innovation type
- Process innovation
- AI technology
- Computer visionPredictive analytics and ML
- Adoption
- Built in-house
- Evidence
- Vendor case study
Sources
Source summaries
A project post on the Dataiku community forum lists the components built for gel detection and classification and gives the status as in progress. It gives estimated savings of 30% of lab analysis time and one FTE of development time.
Full details are in the original sources.