Fig 1.
Fish freshness dataset sample images (a) anchovy (b) horse mackerel.
Fig 2.
Proposed hybrid model structure.
Table 1.
Image counts of the training, testing, and validation dataset.
Table 2.
Models training parameters.
Table 3.
Yolo-v5 loss and avg loss values.
Fig 3.
(a) Training-test accuracy, (b) Loss plots of models using Dataset1.
Fig 4.
(a) Training-test accuracy, (b) Loss plots of models using Dataset2.
Table 4.
Test results obtained from model structures using the training-test dataset.
Fig 5.
Confusion matrix results from (a) Inception-ResNet-v2, (b) Xception model structures using the validation dataset in Dataset1.
Fig 6.
Confusion matrix results from (a) Inception-ResNet-v2, (b) Xception model structures using the validation dataset in Dataset2.
Table 5.
Test results from model structures using the validation dataset.
Fig 7.
Proposed hybrid model structure performance diagram.
Table 6.
Performance test results of the proposed hybrid model.
Fig 8.
Sample images of the performance test (a) anchovy, (b) horse mackerel.
Table 7.
Comparison of proposed hybrid model constructs with existing methods to detect fish freshness.