In our final project for EE 456: Introduction to Neural Networks, we (Alexander Pretka, Shao-Ju Wang, and Ankit Garikipati) explored the capabilities of AlexNet on a fish market dataset. AlexNet is an image classification architecture created by Alex Krizhevsky, Ilya Sutskever, and
Geoffrey Hinton that played a pivotal role in the advancement of deep learning and computer vision.

Our dataset consists of nine different species, each with 1000 images. One unique feature about the dataset is the inclusion of rotated variants of the initial images. While never truly rotation invariant, classification architectures are ideally resilient to rotated images.

свернуть