Fig. 1
Research object and methodology
Fig. 2
Flow chart of clothing image classification
Fig. 3
Schematic diagram of HOG+SVM algorithm
Fig. 4
Architecture of a simple NN
Fig. 5
Architecture of the CNN
Fig. 6
Architecture of small VGG
Fig. 7
Experimental flow chart
Fig. 8
Examples and categories of the Fashion-MNIST dataset
Table 1
Fashion-MNIST dataset label description
| Description | T-shirt/top | Trouser | Pullover | Dress | Notes |
|---|---|---|---|---|---|
| Label | 0 | 1 | 2 | 3 | 4 |
| Description | Sandal | Shirt | Sneaker | Bag | Ankle boot |
| Label | 5 | 6 | 7 | 8 | 9 |
Fig. 9
Accuracy comparison of HOG+SVM, NN, and CNN.
Table 2
Highest accuracy comparison of different models
| Name | Highest Accuracy |
|---|---|
| ML (rbf_4x4) | 91.3% |
| CNN | 89.7% |
| ML (Linear_4x4) | 88.9% |
| NN | 87.7% |
| ML (rbf_8x8) | 86.7% |
| ML (Linear_8x8) | 83.1% |
Fig. 10
Example images of the Fashion144k dataset
Fig. 11
Example images of the SmallV1 dataset
Fig. 12
Recognition accuracy of HOG+SVM and Small VGG for the SmallV1 dataset
Fig. 13
Recognition accuracy of Small VGG network models in different datasets
Fig. 14
Recognition accuracy of the GhostNet model for different datasets
Fig. 15
Accuracy of Small VGG and GhostNet for different numbers of datasets