Computer Vision Specialist
European certification validating mastery of computer vision techniques, from image processing to the implementation of deep learning models for visual analysis.
Key skills
- ✓Image processing and analysis with OpenCV and PIL
- ✓Design and training of CNN for classification and detection
- ✓Implementation of segmentation and object tracking models
- ✓Deployment of vision pipelines in production
The question bank for this exam is prepared after your purchase. The exam is available no later than 48 hours after payment; you will receive an email as soon as it opens.
Is this right for you?
✓ This certification is right for you if:
- →Machine learning engineers specialising in vision
- →Data scientists working on image analysis
- →AI developers transitioning to computer vision
✗ This certification is not suitable if:
- →Beginners without a foundation in machine learning: first orient yourself towards a fundamental ML training before tackling vision.
- →Experts seeking a research-level validation: consider instead the PROVA Computer Vision Architect (Master 801) with a focus on innovative architectures.
Prerequisites & eligibility
Free online registration, prior payment of the exam voucher.
Purchase options
Exam voucher, LMS pathway, bundle packs — build your order
Options d'achat
Composez votre parcours — voucher d'examen, préparation LMS, packs entreprise.
Skills assessed
Advanced CNN architectures
Mastery of ResNet, YOLO, U-Net and visual transformers for real-world use cases.
Image processing
Expertise in preprocessing, data augmentation and feature extraction with OpenCV.
Production deployment
Optimisation and production deployment of models with TensorRT, ONNX and edge computing.
Performance evaluation
Rigorous analysis of metrics (mAP, IoU, F1) and debugging of visual models.
Exam format
Curriculum
Fundamentals of image processing
Colour spaces (RGB, HSV, LAB), pixel operations, spatial filtering (mean, Gaussian, median, bilateral), mathematical morphology (erosion, dilation, opening, closing), edge detection (Sobel, Canny, Laplacian), geometric transformations (rotation, scaling, warping), histograms and equalisation, image pyramids, feature detection (SIFT, SURF, ORB).
Convolutional network architectures
Principles of convolutional, pooling and fully connected layers. Historical architectures: LeNet, AlexNet, VGG. Modern architectures: ResNet and skip connections, Inception and multiscale convolutions, MobileNet and depthwise separable convolutions, EfficientNet and compound scaling. Vision Transformers (ViT) and self-attention. Transfer learning, fine-tuning, feature extraction. Data augmentation and regularisation.
Object detection and segmentation
Two-stage detection: R-CNN, Fast R-CNN, Faster R-CNN with RPN. Single-shot detection: YOLO (v3 to v8), SSD, RetinaNet and focal loss. Detection metrics: IoU, precision, recall, mAP. Non-Maximum Suppression. Semantic segmentation: FCN, U-Net, DeepLab with atrous convolutions. Instance segmentation: Mask R-CNN. Panoptic segmentation. Management of class imbalances.
Advanced techniques and deployment
Object tracking: SORT, DeepSORT, ByteTrack. Pose estimation with OpenPose and MediaPipe. OCR with Tesseract and CRNN. GANs for generation and image augmentation. Model optimisation: quantization (int8, fp16), pruning, knowledge distillation. Deployment formats: ONNX, TensorRT, OpenVINO. Inference on edge devices (Jetson, Coral). Monitoring and drift detection in production.
Tools and frameworks
OpenCV for image processing and classical vision. PIL/Pillow for image manipulation. TensorFlow/Keras and PyTorch for deep learning. Specialised libraries: torchvision, Detectron2, MMDetection, Ultralytics. Dataset management: COCO, Pascal VOC, ImageNet. Annotation tools: LabelImg, CVAT, Label Studio. MLOps for vision: dataset versioning, experiment tracking with Weights & Biases.
How to fund your certification
3 ways to cover the cost of your certification
OPCO — Employer funding
Ask your employer to cover the cost through your OPCO (employee training plan).
Personal — €450 incl. VAT
Secure online payment by credit card (Stripe).
Business rate: €585 excl. VAT
Business — custom quote
Teams, cohorts or private catalogue: a quote tailored to your volumes and goals.
Frequently asked questions
Which frameworks are covered in the exam?
The exam mainly covers PyTorch and TensorFlow for deep learning, OpenCV for image processing, as well as deployment tools like ONNX and TensorRT. Questions are framework-agnostic where possible, favouring understanding of concepts.
Do I code during the exam?
You will not code from scratch but will need to identify errors, optimise pipelines and choose the right architectural approaches. An annotated coding environment is provided.
Does this certification cover deployment in production?
Yes, 20% of the syllabus covers optimisation and deployment: quantization, pruning, ONNX conversion, inference with TensorRT, deployment on edge devices, model monitoring in production and drift management.
Does the PROVA DNA report include my weak points?
Absolutely. The report details your scores by domain (preprocessing, CNN architectures, detection, segmentation, deployment) with precise resource recommendations for improving your areas of weakness, even upon passing.
Your certification pathway
Before, during, after — the recommended progression
Droits du candidat
Transparence
Les critères de décision et le barème sont documentés et accessibles avant l'examen.
Droit d'appel
Toute décision peut être contestée dans les 30 jours. Examen par un tiers indépendant.
Plainte
Toute personne peut signaler un dysfonctionnement. Formulaire public accessible sans compte.
