Digital Poster Session - Analysis: Segmentation
Analysis Methods
Exhibition Hall (Hall 403)
Monday, 06 May 2024 16:00 - 17:00
Spatial-temporal segmentation of cine cardiac MRI time-series
Yingqi Qin1, Fumin Guo1, and Xin Zhou2
1Wuhan National Laboratory for Optoelectronics, Huazhong University of Science and Technology, Wuhan, China, 2State Key Laboratory of Magnetic Resonance and Atomic and Molecular Physics, National Center for Magnetic Resonance in Wuhan, Innovation Academy for Precision Measurement Science and Technology, Chinese Academy of Sciences, Wuhan, China
The proposed method made effective use of the spatial-temporal information in a cine time-series and yielded highly accurate and precise segmentation and cardiac functional measurements, suggesting the utility of our approach for clinical cardiac patient care.
Figure 1. Architecture of the proposed cine time-series segmentation network. 2D cine images from T=20 consecutive cardiac phases were entered into a U-net to extract spatial features for each phase. The extracted spatial features across all the 20 phases were processed by an LSTM/GRU unit. The output was convolved and concatenated with that unsampled from the previous level to serve as input to the next level.
Figure 2. (A) Representative segmentation of basal, mid-cavity, and apical cine at different phases (phase=0, 6, 12, 18, 24) using the proposed algorithm. (B) The volume of LVC, Myo, and RVC provided by our approach (solid lines) and a U-net (dashed lines) for an entire cardiac cycle for one subject. Solid dots indicate LVC, Myo, and RVC measurements derived from manual segmentation at the ED (phase=0) and ES (phase=12).
