TY - JOUR
AU - Summers, Ronald M.
AB - [Radiologists in their daily work routinely find and annotate significant abnormalitiesAbnormalities on a large number of radiology imagesRadiology image. Such abnormalitiesAbnormalities, or lesions, have collected over years and stored in hospitals’ picture archiving and communication systemsPicture Archiving and Communication Systems (PACS). However, they are basically unsorted and lack semantic annotations like type and location. In this paper, we aim to organize and explore them by learning a deep feature representation for each lesion. A large-scaleLarge-scale and comprehensive dataset, DeepLesion, is introduced for this task. DeepLesion contains bounding boxesBounding box and size measurements of over 32K lesions. To model their similarity relationship, we leverage multiple supervision information including types, self-supervised location coordinates, and sizes. They require little manual annotationManual annotation effort but describe useful attributes of the lesions. Then, a triplet network is utilized to learn lesion embeddings with a sequential sampling strategy to depict their hierarchical similarity structure. Experiments show promising qualitative and quantitative results on lesion retrieval, clusteringClustering, and classification. The learned embeddings can be further employed to build a lesion graph for various clinically useful applications. An algorithm for intra-patient lesion matching is proposed and validated with experiments.]
TI - Deep Learning and Convolutional Neural Networks for Medical Imaging and Clinical Informatics: Deep Lesion Graph in the Wild: Relationship Learning and Organization of Significant Radiology Image Findings in a Diverse Large-Scale Lesion Database
DA - 2019-09-20
UR - https://www.deepdyve.com/lp/springer-journals/deep-learning-and-convolutional-neural-networks-for-medical-imaging-A2nYMERgL1
DP - DeepDyve
ER -