Fernando Fernández Salvador, Iván Martín Fernández - MUTSC ETSIT UPM 2023
The goal of the project is to design a classification model using the "Platypus / Perry the Platypus" Kaggle dataset: https://www.kaggle.com/datasets/riyaelizashaju/platypus-or-perry-the-platypus
It should be able to distinguish between pictures of the cartoon character Perry the Platypus when wearing a hat (Perry) and when not wearing it (a nameless Platypus). The main challenge is the reduced size of the dataset, since there are only 9 examples of each class. Therefore, data augmentation techniques will be required.
This repository will be holding all the necessary code as well as the different documents and slides.
- Week 1 Data collection, EDA and data augmentation.
- Week 2 Feature Engineering and final dataset preparation.
- Week 3 Model design and validation. The model selection will be highly influenced by the extracted features.
- Week 4 Model evaluation. If enough time, further model exploration (Deep Learning: YOLO, Transformers etc.).