Dynamic U-shaped convolutional network for mouse cardiac image segmentation and quantification - Takeaways - MDSpire

Dynamic U-shaped convolutional network for mouse cardiac image segmentation and quantification

  • By

  • Yu Wang

  • Wenwen Zhang

  • Wanjun Zhang

  • Cenbin Huang

  • Ming Zhang

  • Naian Xiao

  • Shengge Xu

  • May 28, 2026

  • 0 min

Share

  • 1

    The study introduces a Dynamic U-shaped Convolutional Network (DUCNet) for segmenting mouse cardiac images with myocardial infarction.

  • 2

    DUCNet employs a dynamic convolution to focus on irregular U-shaped structures, enhancing segmentation performance.

  • 3

    A dual-stream fusion block is designed to improve feature extraction and address challenges of small target areas in images.

  • 4

    The proposed method achieved an average Dice coefficient of 80.68%, outperforming existing segmentation models.

  • 5

    A dataset of 243 annotated mouse cardiac slice images is constructed, marking the first of its kind for this research area.

Original Source(s)

Dynamic U-shaped convolutional network for mouse cardiac image segmentation and quantification

  • by Yu Wang, Wenwen Zhang, Wanjun Zhang, Cenbin Huang, Ming Zhang, Naian Xiao, Shengge Xu

  • May 28, 2026

Related Content

Residual inflammatory risk post STEMI: high prevalence despite LDL-C control and association with other secondary prevention targets

Genetics of supraventricular tachycardia: current evidence with a focus on translational relevance and personalized medicine

Correlation between cardiorespiratory fitness and body composition in individuals with different glucose metabolism statuses