ComputerVisionClassification

In this report, there are four models for image classification, using two CNNs, pre-trained fine-tuned AlexNet and DAG-SVM trained of feature extracted from AelxNet.

1. Dataset: Download the 15-scene dataset by link in the paper [Lazebnik et al., 2006].
2. Structure: Ensure images are organized into train/ and test/ folders with subfolders for each class.
3. Model Parameters: Download pretrained AlexNet weights by simply run the code the first time. Those will be automatically downloaded.

On the Jupyter Notebook there are all the steps to run the four models.

Used ChatGPT, free version, for review the grammar in the report and for debug fixing on code (load dataset from folders, load pre-trained model, DAG SVM implementation).