Table 2 DeepInsight-DeepMapper comparison. (*Note: DeepInsight uses 224*224 image resolution, but this resolution is only required if a pre-trained network is utilized).
From: Novel applications of Convolutional Neural Networks in the age of Transformers
DeepInsight | DeepMapper | |
|---|---|---|
Image conversion time | I min 34 s | 334 ms |
Image size | 224 × 224* | 139 × 139 |
Training time | 23 min 37 s | 8 min 56 s |
Accuracy | Train: 1.00 Test: 0.976 | Train: 1.00 Test: 0.975 |
