Pharmaceutical Drugs and Vitamins Synthetic Images
Synthetic dataset of images of drugs and vitamins, organized into 10 classes, useful for visual classification in machine learning.
20,000 JPEG images, 10 classes, COCO annotations, approx. 256 MB
CC BY-SA 3.0
Description
The dataset Pharmaceutical Drugs and Vitamins Synthetic Images contains approximately 20,000 synthetic images depicting pills and supplements commonly distributed in the Philippines. Divided into 10 classes, the images are designed for visual classification tasks, with COCO annotations allowing use in advanced computer vision pipelines.
What is this dataset for?
- Train classification models of medical images of pill
- Test detection algorithms on synthetic data before real use
- Create educational applications about medication
Can it be enriched or improved?
Yes. It is possible to increase the data with transformation techniques (zoom, rotation, brightness), to cross with real databases for a mixed approach, or to integrate more detailed labels (dosage, color, shape). The dataset could also be used to test multi-category analysis models via COCO annotations.
🔎 In summary
🧠 Recommended for
- Computer vision students
- Tests on synthetic data
- Digital health educational projects
🔧 Compatible tools
- PyTorch
- TensorFlow
- FastAI
- Roboflow
- Ultralytics Yolov8
💡 Tip
Use this dataset as a pre-step before moving on to validated medical datasets for serious clinical or commercial uses.
Frequently Asked Questions
Does this dataset contain real or synthetic images?
These are synthetic images generated to simulate pills and vitamins, useful for educational and technical purposes.
Can this dataset be used for a medical project in production?
No, it is intended only for educational or exploratory uses, without official medical validation.
Do COCO annotations allow detection or just classification?
Annotations allow for both: supervised classification and localization for detector models like YOLO or Faster R-CNN.
