Research Article
A New Algorithm for Sketch-Based Fashion Image Retrieval Based on Cross-Domain Transformation
Table 1
The comparison of our dataset with existing datasets.
| Single-modal datasets | Number of images |
| DDAN [28] | 341,021 photos | WTBI [4] | 78,958 photos | DeepFashion [2] | Over 800,000 photos |
| Multi-modal datasets | Number of images |
| QMUL-shoe [8] | 419 sketches, 419 photos | Our Fashion Image dataset | 36,074 sketches, 36,074 photos |
|
|