InternshipKey terms and common mistakes
The words you will hear in Computer Vision work and the mistakes that most often cost interns approvals.
Key terms
| Term | Meaning |
|---|---|
| Pixel | Smallest unit of an image |
| OpenCV | Open-source computer vision library |
| Gaussian blur | Smooths image to reduce noise |
| Classification | Assigning a label to an image |
| Bounding box | Rectangle around an object |
| Semantic segmentation | Class label for every pixel |
| Face recognition | Identifying a person from a face |
| Frame | Single image in a video |
Common mistakes to avoid
- Testing on the same images used for training.
- Ignoring the dataset licence and consent for photos of people.
- Reporting only accuracy without a confusion matrix or failure examples.
- Forgetting that OpenCV loads images as BGR, not RGB.
- Not checking speed (frames per second) on the real device.
Want to go deeper? The free Computer Vision learning path has short modules, games and a certificate for each module.