Tablet defect detection: Datasets & Models

Manual visual inspection of tablets is slow, inconsistent between operators, and impossible to sustain at modern line speeds, which means defective units slip through or good batches get over-rejected. In regulated pharma and semiconductor-adjacent manufacturing, missed defects carry compliance and recall risk, while excess scrap eats into yield. A detection model runs continuously without fatigue, applies the same criteria to every tablet, and produces auditable records of what was flagged and why. That combination reduces escape rates and gives quality teams objective data instead of subjective spot checks.

Automated tablet defect detection delivers consistent, round-the-clock inspection that scales with line throughput and integrates with the cameras and reject mechanisms you already have. RF-DETR provides strong accuracy on small, subtle surface defects, and the model can be retrained as new defect types or product formats appear. Detection logs support quality documentation and continuous improvement without adding headcount.

1. Upload an image: Drop any photo into the widget above to try tablet defect detection for free, right on this page.
2. Preview instant results: The model detects chipped tablet, cracked tablet, discolored tablet, surface spot, broken tablet within seconds and returns an annotated image with structured predictions, so you can immediately see how well it works. Adjust the detection classes and re-run until the results fit your scenario.
3. Like what you see?: With a free Roboflow account, you can deploy this workflow with the Serverless Cloud API, customize it in your workspace, or train RF-DETR on your own images for production accuracy.
4. Get started for free: No code or training data required to begin: upload, review the results, and make it part of your stack in a single click.