Cloud-Based Convolution Neural Network Ensembles for Computer Vision Counting

Explore a 26-minute video presentation by Jifan Zhang, a student at Washington University in St. Louis, detailing an innovative approach to achieve a low RMSE score in a computer vision counting competition. Learn about the team's use of ensemble convolution neural networks and their evaluation of various cloud computing platforms for model training. Gain insights into their methodology, including preprocessing, feature engineering, and model development techniques. The presentation also covers the use of checkpoints and includes a Q&A session. Discover how this team of students, including Yang Cheng, Zhanwen Lu, and Ruohan Zhao, successfully competed against 80+ participants using cloud-based CNN ensembles for computer vision counting tasks.

Introduction
Team Introduction
Collab
Preprocessing
Feature Engineering
Model Development
Checkpoints
QA