What is this programme?
The Computer Vision Basics Certificate from ecertificate.pro is an independent online program covering image preprocessing, classification with transfer learning, object detection, annotation quality control and the practical limits of vision models. It runs about 22 hours and ends with an online assessment. The certificate is delivered inside your private ecertificate.pro panel rather than as an e-mail attachment, and each copy carries a unique verification code that can be checked against the ecertificate.pro database, which is the only place that code is recorded. This program is not government-approved, not officially accredited and does not grant a licence, a protected professional title or any job guarantee. It is a record that you completed this specific course of study and passed its assessment, useful as supporting evidence next to a portfolio of your own work.
Who it is for
Developers and data analysts starting in machine learning, QA and operations staff who handle camera or scanned-document images, annotation team leads, and product managers who need to judge what a vision model can realistically deliver.
What you will be able to do
- Build a reproducible preprocessing pipeline: decode, resize with the right interpolation, normalise to dataset mean and standard deviation, and handle EXIF rotation and channel order (RGB versus BGR).
- Fine-tune a pretrained CNN such as ResNet or EfficientNet on a small custom dataset, using a frozen-backbone stage followed by full unfreezing, and diagnose overfitting from the loss curves.
- Choose augmentations that match real capture conditions - flips, random resized crops, colour jitter, motion blur - and recognise which transforms silently corrupt the labels.
- Read a detection report properly: IoU thresholds, precision-recall curves, mAP@0.5 versus mAP@0.5:0.95, confidence cut-offs and non-maximum suppression settings.
- Run an annotation project in CVAT or Label Studio: write a labelling guideline, measure inter-annotator agreement, and convert among COCO JSON, Pascal VOC XML and YOLO TXT formats.
- Export a trained model to ONNX, measure inference latency and memory on the target hardware, and judge when a smaller input resolution outperforms a heavier architecture.
- Name the failure modes before launch - domain shift, class imbalance, occlusion, low light, tiny-object scale - and apply the privacy rules that cover face or vehicle-plate imagery.
Modules
-
1
Image Data, Colour Spaces and File Traps
-
2
Preprocessing and Augmentation Pipelines
-
3
Image Classification with Transfer Learning
-
4
Evaluation That Survives Review
-
5
Object Detection Foundations
-
6
Data Labelling and Annotation Quality
-
7
Inference, Packaging and Running Cost
-
8
Limits, Bias and Responsible Use
How it works
Learning format
Self-paced written material, downloaded from your dashboard. No fixed schedule and no live sessions — you work through it when it suits you.
Material delivery
Everything appears in your dashboard as soon as your payment is confirmed. Files are served only to your signed-in session.
Prerequisites
Working Python (functions, lists, virtual environments) and comfort reading NumPy array shapes. No prior deep learning, calculus or GPU access is required; every exercise runs on a free hosted notebook or an ordinary laptop CPU.
Completion
Work through the eight modules, then pass the final assessment - multiple-choice plus applied questions on evaluation metrics and annotation decisions - with a score of 70% or higher. Two retakes are included, and the certificate with its verification code appears in your ecertificate.pro panel.
Your certificate
Once your completion is confirmed, we prepare your certificate and upload it to your dashboard, normally within 24–48 hours. You receive an e-mail as soon as it is there — you never have to chase it.
Refunds
Because the material is delivered digitally and immediately, purchases are non-refundable once the files have been made available to your account. The full terms are in our Refund Policy.
