In the article, we proposed a novel centerline extraction framework, which combines an end-to-end trainable multi-task fully convolutional network (FCN) with a minimal path extractor. Compared to a traditional minimal path approach, our method improves patient-level success rate of centerline extraction from 54.3% to 88.8%, according to independent human expert review. To the best of our knowledge, the proposed method is the first deep-learning based centerline extraction method that guarantees single-pixel-wide centerline for a complex tree-structured object, which has addressed multiple long-existing challenges of centerline extraction.
Fig. 1: Schematic workflow of DeepCenterline
Fig. 2: The proposed multi-task FCN architecture.
CuraCloud is proud of the achievement that our scientists have accomplished. We are committed to bringing advanced medical image analysis technology into practical clinical use to benefit patients and healthcare providers.
* IPMI is held biennially since 1969 and is widely recognized as a preeminent international forum for presentation of leading-edge research in the medical imaging field. The acceptance rate of IPMI is around 30%.