|Title: Two Summer Internships||Posted: March 13, 2014|
|Company/Institution: National Institutes of Health Clinical Center|
|Location: Bethesda, Maryland, USA|
|Department: Imaging Biomarkers and Computer-Aided Diagnosis Laboratory|
Description: Project 1, Anatomical landmark detection and labeling in 3D CT scans. We plan to build an internal anatomical structure based coordinates which permits us to know where organs and pathologies locate. They can serve as self-normalized anchor positions instead of traditional volume or world coordinates. Tools: 3D appearance and graphical context modeling, Decision/Regression Random Forest, Deep Convolutional Neural Network.
Project 2, Content prediction on clinical key images. We collect more than 100,000 key images on various liver pathologies (CT, MRI slices) from our clinical PACS system, with weakly supervised transcripts associated (which is extracted from RIS clinical report). We try to build a system to predict or even locate up to 10~20 classes of anatomical structures (whatever is clinical meaningful) presented in the key images. This will be one small but concrete step into the big-data era for medical image analysis. Tools: bound-box based object detection, weakly supervised learning, Deep Convolutional Neural Network.
Application Instructions: Basic Requirements: Paid trainee position, only US permanent resident or citizen are eligible to apply. The internship lasts about 10 weeks with some flexibility on starting date.
Desirable Requirements: Senior undergraduate or graduate student with computer vision experience (e.g., class taken, research projects) and in a related discipline, such as computer science, electrical engineering, mathematics. Medical imaging experience is not a must. Good programming and analytic skills. We will try our best to provide detailed mentoring and technical guidance, to make the summer an enjoyable success.
Please send your application to email@example.com or firstname.lastname@example.org