Renal Replacement Therapy

Table 2 Evaluate results (dropout 0.5)

From: Study on deep learning-based detection of viable cell count in dialysis fluid images

Input image type	Model1		Model2
Input image type	Color	Binary	Color	Binary
Loss	0.1379	0.1941	0.0305	0.0883
Accuracy	0.9590	0.9336	0.9851	0.9805
Precision	0.9761	0.9518	0.9851	0.9842
Recall	0.9179	0.9258	0.9851	0.9766
F1-Score	0.9461	0.9386	0.9851	0.9801

The following figures show the Loss, Acciracy, Precision, Recall, and F1-score of a model (Model 1) using VGG-16 as a feature extractor and transition learning, and a model (Model 2) using some of the weights (fixed up to 15 layers) learned by VGG-16. Dropout is 0.5.

Back to article page

ISSN: 2059-1381

Contact us

Submission enquiries: Access here and click Contact Us
General enquiries: info@biomedcentral.com