EURASIP Journal on Image and Video Processing

Table 2 Detail experimental results regarding each Thai patient

From: Automatic kidney segmentation using 2.5D ResUNet and 2.5D DenseUNet for malignant potential analysis in complex renal cyst based on CT images

Model architecture	Dice score per one patient				Averaged dice score
	501 (benign)	502 (cancer)	503 (benign)	504 (cancer)
Google Colab Pro:
2.5D ResUNet (slice stack of 5)	0.8650	0.6591	0.9128	0.3231	0.6900
NVIDIA DGX A100:
2.5D ResUNet (slice stack of 5)	0.7608	0.4480	0.9050	0.8201	0.7335
2.5D DenseUNet (slice stack of 5)	0.9014	0.7546	0.9132	0.9347	0.8760

The best result regarding each Thai patient is highlighted in bold

Back to article page