Dave Cote, M.Sc.
8+ years of experience in data science, business intelligence and delivering actionnable data-driven solutions that address business challenges. Numerous achievements in predictive modeling and Machine Learning with Python.
Currently Data Scientist in the insurance industry.
View My LinkedIn Profile
View My Medium Profile
Computer Vision
Car Damage Detector (Instance Segmentation and Image Classification)
Project description: Automaticly detect damages on a car picture and produce a damage report. The goal is to experiment image segmentation with Mask-RCNN to see the potential and the complexity applied to a car damage detection problem.
1. Data
- Google Car Images (Demo)
- Damages polygons
2. Model
- Mask-RCNN (Instance segmentation of damage)
- VGG-16 (Damage location and severity classification)
3. Web Interface
- Deploy an interactive Streamlit application to upload a car picture and produce a damage report using a deployed model api with flask.