Hidden Damage Recognition
Problem
- Challenges with identifying product damage prior to shipping to consumer.
- Lack of root cause visibility for hidden damages.
- Quality control only conducted via sampling method for the production line at the factory
- Lack of distribution and production line metrics
Action
- Implemented the usage of cameras and deep learning approach for damage detection and classification for dishwashers.
- Provided Reporting & archiving of the complete production and visualization of damage detection results via dashboards.
- Developed an adaptable & scalable solution with deep learning algorithms and cloud technologies
Tool Stack
- AWS, Python, RabbitMQ, PostgreSQL, .NET
70
%Segmentation accuracy
75
%Classification accuracy
100
%Sampling for quality