Ship Detection using Aerial images

This is a semester project for the course "Artificial Intelligence".

Description

Ship detection from a dataset consisting of aerial images. Image processing techniques like Binarization, Histogram Equalization, Unsharping and Smoothing (using different kernels and median Smoothing) were employed which yielded better results and the processed images were tested on UNET and VGG models.

Loss Function

UNET's Loss Function

Custom Model Without DIP	Custom Model With DIP