Graduation thesis from Posts and Telecommunications Institute of Technology (1/2021)
-
A simple web application demo using Flask Framework for the purpose of testing the accuracy of the built model as well as the practicality of the model.
-
Using VGG-16 to train model for feature extracting with small database (link checkpoints: https://drive.google.com/file/d/1R14bw3JLHHb0DjeDrbDHR6nxtuVbtn-d/view?usp=sharing)
-
Using SVM to train model for classification (link checkpoints: https://drive.google.com/file/d/1lW43qnLu-wT-u0cTgQwvX5hLvWoUUwwF/view?usp=sharing)
-
To build the dataset, the project uses part of the images in the "DeepFashion dataset" collected by researchers from The Chinese University of Hong Kong (Liu, Ziwei and Luo, Ping and Qiu, Shi and Wang, Xiaogang and Tang, Xiaou, DeepFashion: Powering Robust Clothes Recognition and Retrieval with Rich Annotations, 2016.)