Table 5.

Results of Altering the Loss Objective during Training. Assessing the influence of altering the loss objective function during training on the segmentation performance of the proposed cGAN and U-Net. A cGAN was trained with ${\mathcal{L}}_{\mathrm{cGAN}}+\mathrm{\lambda }{\mathcal{L}}_{\mathrm{L}2}$ objective and a U-Net with ${\mathcal{L}}_{\mathrm{L}2}$ objective for 50 epochs followed by a further 50 epochs training with ${\mathcal{L}}_{\mathrm{cGAN}}+\mathrm{\lambda }{\mathcal{L}}_{\mathrm{L}1}$ and ${\mathcal{L}}_{\mathrm{L}1}$ objectives, respectively. Segmentation performances are compared with the previously trained cGANs (${\mathcal{L}}_{\mathrm{cGAN}}+\mathrm{\lambda }{\mathcal{L}}_{\mathrm{L}1}\mathrm{and}{\mathcal{L}}_{\mathrm{cGAN}}+\mathrm{\lambda }{\mathcal{L}}_{\mathrm{L}2}$; $\mathrm{\lambda }=100$; 100 epochs) and U-Nets (${\mathcal{L}}_{\mathrm{L}1}\mathrm{and}{\mathcal{L}}_{\mathrm{L}2}$;100 epochs). Highest DSCs achieved for each tissue are in bold.

Altering the Loss Objective during Training Results
cGAN
Network Loss Objective	F Bone	T Bone	P Bone	F Cartilage	T Cartilage	P Cartilage	VM Muscle	GM Muscle	ACL	PCL
${\mathcal{L}}_{cGAN}+\lambda {\mathcal{L}}_{L1}$	0.918 ± 0.011	0.948 ± 0.018	0.928 ± 0.002	0.812 ± 0.002	0.748 ± 0.042	0.863 ± 0.043	0.113 ± 0.085	0.000 ± 0.000	0.577 ± 0.020	0.073 ± 0.103
${\mathcal{L}}_{cGAN}+\lambda {\mathcal{L}}_{L2}$	0.928 ± 0.015	0.939 ± 0.007	0.921 ± 0.022	0.768 ± 0.016	0.752 ± 0.049	0.862 ± 0.039	0.001 ± 0.001	0.000 ± 0.000	0.652 ± 0.094	0.101 ± 0.074
${\mathcal{L}}_{cGAN}+\lambda {\mathcal{L}}_{L2}$ → ${\mathcal{L}}_{cGAN}+\lambda {\mathcal{L}}_{L1}$	0.936 ± 0.007	0.938 ± 0.021	0.884 ± 0.078	0.800 ± 0.021	0.760 ± 0.035	0.855 ± 0.031	0.739 ± 0.010	0.772 ± 0.005	0.115 ± 0.032	0.392 ± 0.128
U-Net
Network Loss Objective	F Bone	T Bone	P Bone	F Cartilage	T Cartilage	P Cartilage	VM Muscle	GM Muscle	ACL	PCL
${\mathcal{L}}_{L1}$	0.972 ± 0.006	0.960 ± 0.001	0.941 ± 0.010	0.886 ± 0.007	0.834 ± 0.010	0.890 ± 0.034	0.000 ± 0.000	0.000 ± 0.000	0.643 ± 0.153	0.641 ± 0.008
${\mathcal{L}}_{L2}$	0.950 ± 0.007	0.957 ± 0.009	0.939 ± 0.003	0.831 ± 0.020	0.723 ± 0.068	0.837 ± 0.051	0.888 ± 0.000	0.881 ± 0.021	0.491 ± 0.136	0.428 ± 0.196
${\mathcal{L}}_{L2}$ → ${\mathcal{L}}_{L1}$	0.970 ± 0.006	0.961 ± 0.007	0.941 ± 0.003	0.869 ± 0.016	0.793 ± 0.021	0.886 ± 0.027	0.914 ± 0.008	0.933 ± 0.010	0.632 ± 0.170	0.567 ± 0.094

Training and testing were performed on the AMROA training and testing datasets, respectively.

DSCs are presented as mean ± standard deviation.

Abbreviations: FB – femoral bone, TB – tibial bone, PB – patellar bone, FC – femoral cartilage, TC – tibial cartilage, PC – patellar cartilage, VM Muscle - vastus medialis muscle, GM Muscle – medial head of gastrocnemius medialis muscle, ACL – anterior cruciate ligament, PCL – posterior cruciate ligament, DSC - Sørensen–Dice similarity coefficient.