Abstract:
|
Statisticians are considered as experts of data, but are generally only comfortable and familiar with structured data, such as patient outcome and demographics. In the real world, there is a large portion of data that is unstructured. Image is a typical example and is an important data type in the pharmaceutical industry. It is frequently used to diagnose the disease and decide the patient enrollment. Most common statistical methods do not work well with image data unless the image can be transformed to pre-defined features. To the contrary, deep learning can directly learn from image and has become the state-of-the-art algorithm for image analysis. In this presentation, we will introduce the application of deep learning in medical image data analysis and its potential utility in treatment decision and patient selection. We will also include a pilot real world example to show its performance.
|