Car Key-Points Detection Using MobileNet v1
Project information
- Project Name: Car Keypoints detection
- Category: DeepLearning
- Project date: Feb, 2020
- Project URL:
Summary
- Aim of this project is to predict key-points of windshield and headlights for a given car image.
- MobilNet v1 with imageNet weight is used as backbone network with few additional layers. All layers are set to trained.
- Number of epochs trained on are 20.
- Number of model parameters are 24 million.
- Number of images trained on 90, with augmentation 9000.
