- Research Name Vehicle Content Identification via Deep Learning Vision Technology
- Client BIMCON, Inc.
- Category PRACTICUM PROJECTS
- RESEARCH YEAR 2019
The aim of this project was to develop the ability to identify the make of a vehicle by using only images of its interior features and to detect key components of a vehicle’s interior, i.e. identifying the make of a vehicle which is generally known as image classification.
The deep learning algorithm used to develop this model is a Convolutional Neural Network (CNN).
The Convolutional Neural Network architectures used for image classification are EfficientNet for binary classification and MobileNet for multi-class classification. Detecting key components of a vehicle falls under Object Detection within deep learning and it also relies on CNN.TensorFlow, a python-based library that allows us to deploy these models in a python environment, was used for both classification and object detection throughout this project.
Dante Burch, Stephanie Rogers, Jatin Gongiwala, and Ranjitha Vidyashankar