Table 4. The comparison between the two-stage and the end-to-end approaches.

	Two-stage	End-to-end
Training time to obtain stable model (hours)	0.25	1.00
Dedicated GPU memory (GB)	1.9	6.9
Bregma mean error (pixels)	10.12	13.00
Lambda mean error (pixels)	10.15	13.47
Bregma max error (pixels)	30.81	45.00
Lambda max error (pixels)	22.00	43.42