This document discusses using a convolutional neural network called ResNet with transfer learning to automatically detect ships in satellite images. ResNet is a deep residual network that can increase performance efficiency and reduce overfitting. The objective is to detect ships with high accuracy. Transfer learning is used to take a pre-trained ResNet-50 model and reuse it to classify satellite images into ship and non-ship categories, improving accuracy over traditional methods. Convolutional neural networks like ResNet are well-suited for this task as they can learn patterns directly from images without manual feature extraction.