Skip to main content

Table 1 The average values of DSC and ASD with respect to the segmentation performance of 11 models on the validation data and testing data

From: Using 2D U-Net convolutional neural networks for automatic acetabular and proximal femur segmentation of hip MRI images and morphological quantification: a preliminary study in DDH

   

M-1

M-2

M-3

M-4

M-5

M-6

M-7

M-8

M-9

M-10

M-11

Validation data

DSC

Acetabulum

0.9135

0.916

0.9201

0.923

0.9121

0.9281

0.9225

0.9291

0.912

0.9357

0.8837

Proximal femur

0.9161

0.9345

0.9432

0.9196

0.9303

0.942

0.9507

0.948

0.927

0.9474

0.9461

Average

0.9148

0.9253

0.9316

0.9213

0.9212

0.9351

0.9366

0.9386

0.9195

0.9415

0.9149

ASD/mm

Acetabulum

0.701

0.2348

0.2445

0.1176

0.3204

0.4448

0.2268

0.3851

0.4632

0.2418

0.5743

Proximal femur

0.4328

0.2509

0.1919

0.2785

0.2172

0.2692

0.1674

0.1797

0.7393

0.2624

0.2031

Average

0.5669

0.2429

0.2182

0.198

0.2688

0.357

0.1971

0.2824

0.6012

0.2521

0.3887

Testing data

DSC

Acetabulum

0.9096

0.9122

0.9017

0.9087

0.9108

0.9136

0.9117

0.9132

0.9099

0.9148

0.9133

Proximal femur

0.9267

0.923

0.9265

0.9279

0.9226

0.9278

0.923

0.9235

0.9225

0.9238

0.9211

Average

0.9182

0.9176

0.9141

0.9183

0.9167

0.9207

0.9173

0.9184

0.9162

0.9192

0.9172

ASD/mm

Acetabulum

0.3433

0.2776

0.34

0.4047

0.3899

0.2112

0.3955

0.2482

0.1877

0.2226

0.4553

Proximal femur

0.7098

0.4579

0.7855

0.6167

0.6029

0.7436

0.6737

0.5176

0.638

0.543

0.7458

Average

0.5266

0.3677

0.5628

0.5107

0.4964

0.4774

0.5346

0.3829

0.4129

0.3828

0.6006

  1. The values with bold italics indicated that the optimal performance was carried out by one model, as well as the larger DSC and the smaller ASD values indicated the better segmentation
Back to article page

BioMedical Engineering OnLine

ISSN: 1475-925X

Contact us