Introduction

In this project, we implemented the recognition of five types of flowers which are daisy, dandelion, rose, sunflower and tulip. By using the base line model, the test accuracy is close to 71.3% and getting overfitting. Through our effort, the final accuracy can reach 92.6% by adopting the deeper network resnet18.

Details

Subject: Computer Vision

Year: 2022

Student Name: Lujie Ma, Wenlong Dai

Student ID: a1810558, a1777081

Comptetion Name: Flowers Classification Competition

Final Results: ACC: 92.6% FLOPs: 3.59G