A ViTUNeT-based model using YOLOv8 for efficient LVNC diagnosis and automatic cleaning of dataset

Figures

• 

  Figure 1:. Segmentation of the left ventricle, illustrating the external layer (green), the inner cavity (blue), and the trabeculae or non-compacted area (yellow).

• 

  Figure 2:. Comparison between the number of patients and MRI slices in each group. (a) Numbers of patients in each group. (b) Numbers of MRI slices in each group.

• 

  Figure 3:. OpenCV library is used to find the outermost contour in segmentation masks.

• 

  Figure 4:. Performance of the YOLOv8m model trained on the cardiology dataset.

• 

  Figure 5:. Examples of slices that cause issues for the models. (a) MRI slice no ring. (b) MRI slice showing LV with small area and artefacts.

• 

  Figure 6:. Detections made by model opening-IC-LV on slices of the No Ring class.

• 

  Figure 7:. Relationship between Conf-Valid-IC and Conf-Open-IC, and the linear model (red line) that best distinguishes No Ring and poor-quality cuts (low Conf-Valid-IC value) from the rest.

• 

  Figure 8:. Relationship between values Conf − Valid − IC and Area − bbox − Valid − IC of the MRI slices.

• 

  Figure 9:. Flowchart establishing a criterion for automatically discarding slices from the dataset.

• 

  Figure 10:. Number of MRI slices to be discarded for each patient group according to the flowchart in Figure 9.

• 

  Figure 11:. ViTUNeT architecture. Inspired by the TransUNet architecture [18].

• 

  Figure 12:. Integration of the YOLOv8 model with the ViTUNeT model.