This document discusses classifying chest pathology images using deep learning techniques. It explores using pre-trained convolutional neural networks (CNNs) to classify chest radiograph images as either healthy or pathological, and to identify specific pathologies. The document reviews previous work on applying deep learning to medical image analysis. It then proposes using features extracted from pre-trained CNN models to classify chest radiographs, focusing on classifying images as healthy vs. pathological as an important screening task. The strengths of deep learning approaches for analyzing various chest diseases are explored.