We believe the best dataset is even more compelling than the best algorithm.
Deep learning is a new area of machine learning research which advances us towards the goal of artificial intelligence. Deep learning includes multiple levels of representation and abstraction to make sense of data such as images, sound, and text.
Many internal cancers go undetected due to a lack of symptoms in early stages. Using AI we are developing techniques for the 3D reconstruction, detection, and tracking of internal growths over time, in the hopes of providing in-home monitoring of cancers.
This site uses three machine learning tasks (identification, segmentation and classification) to collect relevant data from abdominal, vascular, thyroid and baby brain ultrasound scans.
The goal of this website is to create the largest and most meaningful dataset of ultrasound images. We created three challenges to collect data: Classify, Identify, and Segment
You will view a set of images from an ultrasound study of the abdomen, baby brain, thyroid or vascular system. Your challenge is to flag those images that show abnormalities.
You will view an image from an abdominal ultrasound study. Your challenge is to select the names of all the organs that are visible on the ultrasound image.
You will view an image from an abdominal ultrasound study. Your challenge is to use our online drawing tool to outline specific organs on the image.
Once completed, verified and annotated, this dataset will be made publicly available to the research community.
Based on points awarded for accuracy and activity, the top three participants will be offered co-authorship roles on deep learning publications related to this dataset.
Not an expert? Read our free tutorials and get evaluation help from Stanford University and Hospital doctors on how to properly examine ultrasound images and increase your scores on our challenges.Go To Tutorials
Most of our code is done in Python. From the data de-identification to the Convolution Neural Networks implementation and training (tensorflow library), including images and dicom pre-processing. All of this will be publicly available to download or on Github.
Our datasets contain for now the following categories of scans listed below.
Once fully classified, each of those datasets will be made partially available to the public for research purposes.
Currently a Graduate Student and Research Assistant in the Artificial Intelligence department of Stanford University, Alexandre is advised by Sebastian Thrun on this project for Deep learning and Ultrasound analysis. Working in the fields of artificial intelligence, machine learning, and deep learning for several years, his current research interests are to broadly applying them to domains such as healthcare, and medicine.
Leandra is a PhD candidate in the Electrical Engineering Department. With a background in optics, light transport and fabrication, recent research focuses on image processing and deep learning of ultrasound images and volumes under the supervision of Dr. Jeremy Dahl. Research interests include biomedical applications of machine learning using deep learning and reinforcement learning.
Professor Sebastian Thrun pursues research on robotics, artificial intelligence, education, human computer interaction, and medical devices. His work with Andre Esteva and Brett Kuprel, which produced outstanding results on deep learning and skin cancer detection, was recently published on the cover of Nature. As a founding member of Google X, Thrun pioneered innovative projects like Google Glass. He won numerous awards, including the Max Planck Research Award. Professor Thrun’s role in this project includes subject matter expertise and deep learning oversight.
Dr. Halabi is a Clinical Assistant Professor at the Stanford University School of Medicine and Medical Director for Radiology Informatics at Stanford Children's Health. He is a practicing fetal and pediatric radiologist at Lucile Packard Children's Hospital. His current areas of academic and research interest include: imaging informatics, deep/machine learning in imaging, artificial intelligence in medicine, clinical decision support, enterprise imaging, outside image management and patient-centered care. Dr. Halabi's role in this project includes data and image curation, subject matter expertise and physician oversight.
Dr. Muelly is a radiology fellow and clinical instructor at Stanford University School of Medicine. He is currently completing a fellowship in body MRI and completed his residency training at Stanford, as well. His recent research interests include the application of deep and machine learning techniques to radiology. Dr. Muelly's role in this project includes data and image curation, subject matter expertise and physician oversight.
Ali Thabet is a Research Scientist at King Abdullah University of Science and Technology (KAUST). His research interests are in applications of AI, specifically deep learning, in the field of computer vision. He is also a pationate developer. His primary role in this project includes frontend and backend development, as well as implementation of some deep learning algorithms.
please call us on
Cell: (999) 123 4567
Fax: (999) 123 4567
City, ST ZIP