Table of Contents
Computer Vision
A Cross-Vendor Training Guide
Certification: AWS ML Specialty, Azure AI-102, Google ML Engineer, CompTIA AI+
Introduction
Computer Vision enables machines to interpret and understand visual information from images and videos. From facial recognition to autonomous vehicles, CV powers some of the most impactful AI applications today.
Core CV Tasks
|
Task |
Description |
Examples |
|
Image Classification |
Assign label to entire image |
Cat vs dog, product category |
|
Object Detection |
Locate and classify objects |
Pedestrians, vehicles, defects |
|
Semantic Segmentation |
Label each pixel by class |
Road scenes, medical imaging |
|
Instance Segmentation |
Separate individual objects |
Count items, robotics |
|
Pose Estimation |
Detect body/object keypoints |
Sports analysis, AR/VR |
|
OCR |
Extract text from images |
Document processing |
|
Face Recognition |
Identify or verify faces |
Security, authentication |
Image Fundamentals
Digital Image Representation
- Pixels: Basic unit, grid of values
- Channels: RGB (3), Grayscale (1), RGBA (4)
- Resolution: Width × Height in pixels
- Bit Depth: Values per channel (8-bit = 0-255)
Image Preprocessing
- Resizing: Scale to model input size
- Normalization: Scale pixel values (0-1 or -1 to 1)
- Augmentation: Flip, rotate, crop, color jitter
- Color Space: RGB, BGR, HSV, Grayscale conversion
Convolutional Neural Networks
CNNs are the foundation of modern computer vision, designed to process spatial data.
Key Layers
|
Layer |
Purpose |
|
Convolutional |
Extract features using learnable filters (edges, textures, patterns) |
|
Pooling |
Downsample, reduce computation, add invariance (Max, Avg) |
|
Batch Normalization |
Stabilize training, allow higher learning rates |
|
Dropout |
Regularization, prevent overfitting |
|
Fully Connected |
Final classification layers |
CNN Architectures
|
Model |
Year |
Key Innovation |
|
AlexNet |
2012 |
Deep CNN, ReLU, dropout, GPU training |
|
VGGNet |
2014 |
Deeper (16-19 layers), 3×3 filters only |
|
GoogLeNet/Inception |
2014 |
Inception modules, multiple filter sizes |
|
ResNet |
2015 |
Skip connections, very deep (152+ layers) |
|
EfficientNet |
2019 |
Compound scaling, efficiency |
|
Vision Transformer |
2020 |
Transformers for images, patches as tokens |
Object Detection
Locate objects in images with bounding boxes and class labels.
Detection Architectures
|
Model |
Type |
Characteristics |
|
R-CNN Family |
Two-stage |
Region proposals then classify. Accurate but slower |
|
YOLO |
One-stage |
Single pass, real-time. Fast, good accuracy |
|
SSD |
One-stage |
Multi-scale detection, balance speed/accuracy |
|
DETR |
Transformer |
End-to-end, no anchors needed |
Detection Metrics
- IoU: Intersection over Union (overlap measure)
- mAP: Mean Average Precision across classes
- Precision/Recall: At various IoU thresholds
Image Segmentation
Segmentation Types
|
Type |
Description |
|
Semantic |
Every pixel labeled by class (all cars = same label) |
|
Instance |
Separate each object instance (car1, car2, car3) |
|
Panoptic |
Combines semantic + instance segmentation |
Segmentation Models
- U-Net: Encoder-decoder with skip connections. Medical imaging
- Mask R-CNN: Extends Faster R-CNN with mask branch
- DeepLab: Atrous convolutions, multi-scale
- SAM: Segment Anything Model – foundation model for segmentation
OCR & Document AI
Extract text and structure from images and documents.
OCR Pipeline
- Text Detection: Find text regions
- Text Recognition: Convert regions to text
- Layout Analysis: Understand document structure
- Post-processing: Spell check, formatting
Transfer Learning in CV
Use pre-trained models as starting point for new tasks.
Transfer Learning Strategies
- Feature Extraction: Freeze backbone, train new classifier
- Fine-Tuning: Unfreeze some/all layers, lower learning rate
- Pre-trained Backbones: ImageNet, COCO, OpenImages
Vendor CV Services
|
Vendor |
Service |
Capabilities |
|
AWS |
Amazon Rekognition |
Faces, objects, text, moderation |
|
AWS |
Textract |
Document OCR, forms, tables |
|
|
Cloud Vision API |
Labels, faces, OCR, landmarks |
|
|
Document AI |
Document parsing, extraction |
|
Microsoft |
Azure AI Vision |
Analysis, OCR, spatial analysis |
|
Microsoft |
Document Intelligence |
Forms, invoices, receipts |
Documentation Links
- AWS Rekognition: aws.amazon.com/rekognition/
- Google Vision: google.com/vision/docs
- Azure Vision: microsoft.com/azure/ai-services/computer-vision/
Key Takeaways
- CV enables visual understanding – classification, detection, segmentation
- CNNs are foundational – convolutions, pooling, hierarchical features
- Architecture choice matters – ResNet, EfficientNet, ViT
- Detection vs segmentation – boxes vs pixel-level masks
- Transfer learning is essential – leverage pre-trained models
- Cloud services simplify CV – Rekognition, Vision API, Azure Vision
Resources
- PyTorch Vision: org/vision/
- TensorFlow Hub: dev
- Ultralytics YOLO: ultralytics.com
- Hugging Face Vision: co/docs/transformers/tasks/image_classification
Article 13 | Computer Vision
Level: Intermediate | Reading Time: 25 min | Feb 2025
PowerKram Career Preparation Resources
Preparing for a certification exam aligned with this content? PowerKram offers objective-based practice exams built by industry experts, with detailed explanations for every question and scoring by vendor domain. Start with a free 24-hour trial:
- AWS ML Specialty Practice Tests — Computer vision and Rekognition objectives for the AWS Certified Machine Learning Specialty exam
- Azure AI-102 Practice Tests — Azure AI Vision service objectives for the AI Engineer Associate exam
- Google Cloud ML Engineer Practice Tests — Vision AI objectives for the Google Professional ML Engineer certification
Part of the Complete AI & Machine Learning Guide
This article is part of The Complete Guide to AI and Machine Learning, a comprehensive pillar guide covering every essential AI/ML discipline from foundations to production deployment. The pillar guide maps how this topic connects to the broader AI/ML ecosystem and provides business context, common misconceptions, and underutilized capabilities for each area.
Continue Your Learning
Explore these related articles in the AI/ML training series to deepen your expertise across the full stack:
- Deep Learning and Neural Networks — For the CNN architectures, training techniques, and GPU computing that power computer vision
- Machine Learning Fundamentals — For foundational concepts like classification, evaluation metrics, and the ML workflow
- Model Evaluation and Validation — For detection metrics like mAP, IoU, and precision-recall analysis
- MLOps and Model Deployment — For deploying vision models to edge devices and production endpoints
- Data Preparation and Feature Engineering — For image preprocessing, augmentation pipelines, and feature extraction techniques
- Implementation specialist
← Return to the Complete AI & Machine Learning Guide for the full topic map and all supporting articles.
Question #1
A data science team at a consumer lending company is building an AI model to approve or deny personal loan applications. The compliance officer insists the model must achieve Demographic Parity, Equalized Odds, AND Predictive Parity simultaneously to satisfy all stakeholders. The lead ML engineer pushes back, citing a fundamental limitation.
Why is the compliance officer’s requirement problematic?
A) These three metrics can only be satisfied simultaneously if the model uses protected attributes as direct input features.
B) Achieving all three metrics requires an interpretable model architecture such as logistic regression, which would sacrifice accuracy.
C) These metrics are designed for classification tasks only and cannot be applied to the continuous probability scores used in lending decisions.
D) It is mathematically proven that — except in trivial cases — Demographic Parity, Equalized Odds, and Predictive Parity cannot all be satisfied simultaneously, so the organization must choose which definition of fairness is most appropriate for their context.
Solution
Correct Answer: D
Explanation: This reflects the Impossibility Theorem described in the Fairness Metrics section. These three fairness definitions are mathematically incompatible in all but trivial cases (e.g., when base rates are identical across groups). Organizations must make a deliberate, documented choice about which fairness metric best fits their use case, regulatory requirements, and stakeholder values. The other options introduce incorrect preconditions — using protected attributes, requiring specific architectures, or limiting metric applicability — none of which are the actual constraint.
Question #2
A consortium of five hospitals wants to collaboratively train a diagnostic AI model for a rare disease. Data privacy regulations such as HIPAA prohibit sharing patient records across institutions, and no single hospital has enough data to train an accurate model independently. The consortium needs a technique that enables collaborative model training while keeping all patient data within each hospital’s infrastructure.
Which privacy-preserving technique is BEST suited to this scenario?
A) Homomorphic encryption, which allows the hospitals to upload encrypted patient records to a shared cloud server where the model is trained on ciphertext without ever decrypting the data.
B) Federated learning, where a global model is sent to each hospital, trained locally on that hospital’s patient data, and only aggregated model updates — not raw data — are shared with a central server.
C) Differential privacy, which adds calibrated noise to each hospital’s patient records before they are combined into a single centralized training dataset.
D) Synthetic data generation, where each hospital creates artificial patient records that mimic statistical patterns and then shares the synthetic datasets for centralized model training.
Solution
Correct Answer: B
Explanation: Federated learning is specifically designed for this scenario — it enables collaborative model training across decentralized data sources without centralizing the raw data. The model travels to the data, not the other way around. Each hospital trains locally, and only model gradients (updates) are aggregated centrally. While homomorphic encryption is a valid privacy technique, it is computationally expensive and does not directly address the distributed training challenge. Differential privacy with centralized data still requires sharing records. Synthetic data loses fidelity for rare diseases where subtle clinical patterns matter most.
Question #3
A corporate legal department has deployed an AI system to review vendor contracts and flag potentially risky clauses. After initial deployment as a fully automated system (human-out-of-the-loop), the tool missed several unusual liability clauses that fell outside its training patterns, exposing the company to significant financial risk. Leadership wants to redesign the system to balance efficiency with risk mitigation.
Which approach BEST addresses this situation while maintaining operational efficiency?
A) Retrain the model on a larger dataset of contracts that includes the unusual liability clauses it missed, then redeploy as a fully automated system with quarterly accuracy audits.
B) Replace the AI system entirely with a team of paralegals who manually review all contracts, since AI has proven unreliable for legal document analysis.
C) Implement a human-on-the-loop model with confidence-based routing, where high-confidence contract reviews are auto-approved with sampling, and low-confidence or high-value contracts are escalated to attorneys for review.
D) Switch to an interpretable rule-based system that uses keyword matching to flag risky clauses, since black-box AI models cannot be trusted for legal decisions.
Solution
Correct Answer: C
Explanation: The human-on-the-loop model with confidence-based routing directly addresses the core problem: fully automated systems miss edge cases, while fully manual review is inefficient. By routing decisions based on the model’s confidence level, the organization captures the efficiency benefits of automation for routine contracts while ensuring human expertise is applied to uncertain or high-value cases. This matches the document’s guidance that the appropriate level of human oversight should be calibrated to the risk, impact, and reversibility of decisions. Simply retraining doesn’t prevent future novel patterns from being missed. Abandoning AI entirely sacrifices the efficiency gains. Rule-based keyword matching is too rigid for complex legal language.
Question #4
A fintech company uses a gradient-boosted ensemble model to evaluate personal loan applications. A financial regulator has issued an inquiry requiring the company to provide individual-level explanations for each applicant who was denied credit — specifically, they must cite the top contributing factors for every adverse decision and show applicants what changes would improve their outcome.
Which combination of explainability techniques BEST satisfies both regulatory requirements?
A) SHAP values to identify the top features contributing to each denial, combined with counterfactual explanations to show applicants the smallest changes that would produce a different outcome.
B) Global feature importance rankings to show which factors the model weighs most heavily across all decisions, combined with partial dependence plots to illustrate how each feature affects predictions on average.
C) A global surrogate model (decision tree) trained to approximate the ensemble’s behavior, which can then be presented to regulators as the actual decision logic.
D) Attention visualization to show which parts of the application the model focuses on, combined with LIME to fit a local linear model around each prediction.
Solution
Correct Answer: A
Explanation: The regulator requires two things: (1) individual-level factor attribution for each denial, and (2) actionable guidance for applicants. SHAP values provide mathematically rigorous, game-theoretic feature contributions for individual predictions — making them the gold standard for per-decision explanations. Counterfactual explanations identify the smallest input changes needed to flip the outcome, directly addressing the ‘what would need to change’ requirement. Global feature importance and PDP are aggregate techniques that do not explain individual decisions. A surrogate model is an approximation and misrepresents the actual decision process. Attention visualization applies to neural networks and transformers, not gradient-boosted ensembles.
Question #5
A global consumer brand is deploying a generative AI system to create personalized marketing emails at scale across diverse international markets. During pilot testing, the system occasionally produces culturally insensitive content when targeting specific demographic segments, including stereotypical references and tone-deaf messaging that could damage the brand’s reputation.
Which set of safeguards is MOST comprehensive for responsible deployment of this generative AI system?
A) Translate all marketing content into English first, run it through a single toxicity filter, and then translate it back into the target language before sending.
B) Restrict the generative AI to producing content only in English for all markets, and hire local translators to manually adapt every email for cultural relevance.
C) Add a disclaimer to each email stating that the content was generated by AI, which satisfies transparency requirements and shifts responsibility away from the brand.
D) Implement a multi-layer pipeline: prompt engineering with cultural sensitivity guidelines, automated toxicity and bias detection on outputs, human review sampling with higher rates for diverse segments, and a recipient feedback mechanism to flag inappropriate content.
Solution
Correct Answer: D
Explanation: The multi-layer pipeline approach addresses the problem at every stage — from input (prompt engineering with cultural guidelines), through processing (automated toxicity and bias detection), to output (human review sampling and recipient feedback). This aligns with the document’s guidance on responsible generative AI deployment, which emphasizes content filtering, human review for high-stakes content, transparent disclosure, and red-team testing. Translating to English and back introduces translation artifacts and misses cultural nuance. Restricting to English ignores the reality of global marketing. A disclaimer alone does not prevent the harm — it merely attempts to deflect accountability, which contradicts the core principle of accountability in responsible AI.
Choose Your AI Certification Path
Whether you’re exploring AI on Google Cloud, Azure, Salesforce, AWS, or Databricks, PowerKram gives you vendor‑aligned practice exams built from real exam objectives — not dumps.
Start with a free 24‑hour trial for the vendor that matches your goals.
- All
- AWS
- Microsoft
- DataBricks
- Salesforce




