Erik Termes
-
Description: This project is a computer vision deep learning model able to identify the brand of a pair of sneakers based on a picture.
-
Data: The data and images used have been webscraped from zalando, and the model has been build using deep learning techniques.
-
Features: The project so far can distinguish the brand of a pair of sneakers between Nike or Adidas, the next steps would be to expand the scope of the brands it is able to identify or train it with more clothing items.
-
The model: The model is a deep learning sequential model at around 85% accuracy.
-
The demo: There is an available demo for this project, you can test the model by taking a picture with the computer webcam.
-
Instructions: Due to GitHub file size restictions, the data files have been removed, so the code must be run from start in the following order: Shoe scrapper -> Data cleaning -> Sneakers deep learning model -> Demo shoe recommender - Final Product