Deep Learning for Automatic Pneumonia Detection

Data Science
Air 5

Тезисы

Pneumonia is the leading cause of death among young children and one of the top mortality causes worldwide. The pneumonia detection is usually performed through examine of chest X-Ray radiograph by highly-trained specialists. This process is tedious and often leads to a disagreement between radiologists. Computer-aided diagnostic systems showed the potential for improving diagnostic accuracy. In this work, we develop the computational approach for pneumonia regions detection based on single-shot detectors, squeeze-and-exitation deep convolution neural networks, augmentations and multi-task learning. The proposed approach was evaluated in the context of the Radiological Society of North America Pneumonia Detection Challenge, achieving one of the best results in the challenge.

Аудитория и уровень

Researchers, Data Scientists, Healthcare professionals.

Shenzhen Research Institute of Big Data

Александр Калинин

Dr. Alexandr Kalinin is a PostDoctoral Researcher jointly at the University of Michigan and Shenzhen Institute of Big Data, China. He received his PhD in Bioinformatics from the University of Michigan in 2018. His research focuses on studying cellular organization, detection and diagnosis of various pathological conditions, and predicting personalized treatment response from biomedical imaging data using modeling, visual analytics, and machine and deep learning methods.

Dr. Alexandr Kalinin is a PostDoctoral Researcher jointly at the University of Michigan and Shenzhen Institute of Big Data, China. He received his PhD in Bioinformatics from the University of Michigan in 2018. His research focuses on studying cellular organization, detection and diagnosis of various pathological conditions, and predicting personalized treatment response from biomedical imaging data using modeling, visual analytics, and machine and deep learning methods.