We provide the pytorch implementation of a semantically constrained GAN to generate artificial realisitc fruit images for training to reduce the need of real image labeling in fruit detection.
pip install -r requirements.txt
Use generateGANImages.ipynb notebook to load the pre-trained Semantic Consistent GAN model and generate target domain images from source synthetic image.
The dataset used in this research:
- Synthetic Grape Data
The dataset used in this research:
- Night Grape
- Day Grape