Repository for Group M3's projects. Members: Kairos, James, David, Brayden, See Lun
pip install tensorflow
pip install numpy
There is an increasing amount of waste generated from fast fashion, from 182 000 tons of leather and textile waste disposed in 2021 to 249 000 tons in 2022, and is something to be worried about. Therefore, our group set foot on eXApparel to reduce the amount of clothing waste generated by promoting the idea of sustainable fast fashion.
Encourages users to exchange their unwanted but as good as new clothes at a popup store. The app allows for the user to request for an exchange of clothes and the clothing that the user wants would be prepared by the popup store. Upon arrival at the store, users of eXApparel can exchange the clothes that they brought, piece for a piece as requested from the application. They can upload an image to the application and get sorted which will be added into the database of all the clothing that is stored in the pop-up store. When the user receives the set of clothes from the store, the database of the clothes the store contains would be updated, ready for the next user to patronize the store. The cycle repeats itself for every user, allowing many users to try out different clothing that they do not need to buy, while still enjoying the joy of wearing relatively new clothes everyday.
We built on an existing VGG16 CNN model to identify and classify the images of the clothing that is input into the model. The clothing are sorted according to: 1: Dress 2: Hat 3: Longsleeved 4: Outwear 5: Pants 6: Shirt 7: Shoes 8: Shorts 9: Skirt 10: T-shirt Our training accuracy is at about 83.33 %, while our test accuracy is at about 84.41%.
We had to spend a lot of time (many hours) to train the models on Google Colab, but we still managed to finish our project for submission.
Through this project and hackathon, we learnt a lot more about machine learning and tried hands on to really start a model with what we learnt from the workshops. The workshops and hackathon experience allowed us to gain a lot more knowledge and insight to what machine learning has to offer for us, and realized that it is not as hard as we think it would be. We hope to bring our knowledge of machine learning with us on our learning in future, and work on more interdisciplinary projects that incorporate AI.