What You'll Learn
Prerequisites
- No prior programming experience needed
- Basic computer knowledge
- A laptop with internet access
- Enthusiasm to learn!
Tools & Setup
📖 Tutorial Chapters & Curriculum
Follow the structured learning path from beginner to advanced
1
What is Computer Vision? Use Cases, Skills, and Roadmap
A practical foundation: what CV is, where it’s used, and how to learn it like an engineer.
2
Sampling & Quantization: Resolution, Bit-Depth, and Visual Artifacts
Learn why aliasing and banding happen and how it impacts models.
3
Normalization: Scale, Mean/Std, and Training Stability
Make training stable and avoid serving mismatch.
4
Keypoints & Descriptors: The Core Idea Behind Matching
Why keypoints are stable and how descriptors enable matching.
5
CNN Layers Explained: Convolution, Pooling, Activation, BatchNorm
Understand each building block and its role.
6
VGG, Inception, ResNet: Ideas That Shaped Modern CNNs
Core architectural ideas: depth, multi-branch, and residual learning.
7
Detection Basics: Boxes, Anchors, IoU
Vocabulary for detection systems.
8
Segmentation Types: Semantic vs Instance vs Panoptic
Choose the correct segmentation approach.
9
Why Transfer Learning Works
- Why Transfer Learning Works 16 min
10
ViT Basics: Patches and Tokens
How images become tokens for transformers.
11
Face Recognition Pipeline
- Face Recognition Pipeline 18 min
12
Video Fundamentals
- Video Fundamentals 16 min
13
Classification Metrics
- Classification Metrics 16 min
14
From Notebook to API
- From Notebook to API 17 min
15
GANs vs Diffusion
- GANs vs Diffusion 18 min
16
Images as Tensors: Pixels, Channels, and Batch Dimensions
Understand how images become arrays and what shapes models expect.
17
Color Spaces: RGB vs HSV vs LAB (When to Switch)
Pick color spaces for segmentation, enhancement, and lighting robustness.
18
Resizing Strategies: Aspect Ratio, Padding, Letterboxing
Preserve geometry and avoid distorted objects.
19
HOG Features: Why Gradients Capture Shape
Understand HOG and why it powered early pedestrian detection.
20
Backprop in CNNs: A Practical Intuition
Why gradients vanish/explode and how modern architectures fix it.
21
Residual Connections Deep Dive
Why identity shortcuts stabilize optimization.
22
Two-Stage Detectors: Faster R-CNN
Region proposal + classification for high accuracy.
23
U-Net Architecture
- U-Net Architecture 18 min
24
Feature Extraction vs Fine-Tuning
Freeze backbone or fine-tune for domain adaptation.
25
ViT vs CNN
- ViT vs CNN 18 min
26
Face Detectors
- Face Detectors 18 min
27
Tracking-by-Detection
- Tracking-by-Detection 18 min
28
Confusion Matrix Reading
- Confusion Matrix Reading 15 min
29
Export Formats
- Export Formats 18 min
30
Diffusion Intuition
- Diffusion Intuition 20 min
31
Problem Framing: Classification vs Detection vs Segmentation
Pick the correct CV formulation to avoid wrong architectures and metrics.
32
Convolution Intuition: Kernels, Filters, and Feature Detection
Understand blur, sharpen, Sobel edges—and why CNNs generalize this.
33
Augmentation Fundamentals: What Helps vs What Hurts
Augment for realism and robustness, not randomness.
34
ORB vs SIFT vs SURF: Trade-offs for Real Projects
Pick features based on speed, robustness, and constraints.
35
Training a Classifier: Losses, Optimizers, and Schedules
Train properly with cross-entropy, Adam/SGD, and LR schedules.
36
EfficientNet and Compound Scaling
Scale depth/width/resolution together for best accuracy per FLOP.
37
One-Stage Detectors: YOLO/SSD
Single pass prediction for real-time use.
38
Segmentation Losses
- Segmentation Losses 19 min
39
Domain Shift Problems
- Domain Shift Problems 18 min
40
Hybrid CNN-Transformer Models
Combine efficient local features with global attention.
41
Face Embeddings and Metric Learning
Triplet/ArcFace intuition and representation learning.
42
Optical Flow
- Optical Flow 17 min
43
Detection Metrics: mAP
- Detection Metrics: mAP 18 min
44
TensorRT Optimization
- TensorRT Optimization 19 min
45
Vision-Language Models
- Vision-Language Models 19 min
46
Dataset Strategy: Diversity, Labeling Rules, and Leakage Prevention
Build datasets that survive production—diversity beats raw volume.
47
Histograms & Contrast: Equalization and CLAHE
Improve visibility and handle uneven illumination.
48
MixUp, CutMix, Mosaic: Modern Augmentations Explained
Strong augmentations for robustness (especially detection).
49
Template Matching: High-Speed Detection in Controlled Scenes
When it works and how to harden it.
50
Overfitting: Detection and Fixes That Actually Work
Recognize overfitting and fix it with data and discipline.
51
MobileNet for Edge Vision
- MobileNet for Edge Vision 16 min
52
Non-Max Suppression (NMS)
- Non-Max Suppression (NMS) 16 min
53
Mask R-CNN Overview
- Mask R-CNN Overview 18 min
54
Layer Freezing Recipe
- Layer Freezing Recipe 17 min
55
Swin Transformer
- Swin Transformer 19 min
56
Threshold Tuning: FAR/FRR
- Threshold Tuning: FAR/FRR 18 min
57
3D CNNs for Action Recognition
Temporal kernels and compute trade-offs.
58
Segmentation Metrics
- Segmentation Metrics 17 min
59
Docker for Inference
- Docker for Inference 17 min
60
Self-Supervised Vision
- Self-Supervised Vision 19 min
61
Classical CV vs Deep Learning: A Practical Decision Guide
When OpenCV-style methods still win and how to build hybrid systems.
62
Noise Models & Denoising: Gaussian, Salt-and-Pepper, Motion Blur
Choose denoising methods without removing signal.
63
Augmentation for Detection & Segmentation: Label-Safe Transforms
Keep boxes/masks correct during transforms.
64
Corner Detection: Harris and Shi-Tomasi in Practice
Detect stable features for tracking and geometry tasks.
65
Grad-CAM and Feature Map Visualization
Understand where the model ‘looks’ and catch spurious cues.
66
Normalization Choices in Vision
BatchNorm vs GroupNorm vs LayerNorm in practice.
67
Small Object Detection
- Small Object Detection 18 min
68
Mask Post-Processing
- Mask Post-Processing 16 min
69
Fine-Tuning With Limited Data
Regularization and augmentation best practices.
70
Attention Maps: Interpretability
Visualizing attention safely and understanding limitations.
71
Liveness Detection
- Liveness Detection 20 min
72
Temporal Models: LSTM vs Transformers
Sequence modeling over per-frame features.
73
Calibration
- Calibration 18 min
74
Batching vs Real-Time
- Batching vs Real-Time 16 min
75
Synthetic Data
- Synthetic Data 18 min
76
CV System Design Basics: Latency, Accuracy, and Cost
Architect-level trade-offs for real systems.
77
Geometric Transforms: Affine vs Perspective (Homography)
Warp images correctly for docs, lanes, and planar scenes.
78
Lighting Robustness: Color Augmentations That Simulate Reality
Handle device differences and exposure shifts.
79
Optical Flow: Motion Tracking Between Frames
Lucas–Kanade intuition and limitations.
80
Batch Size vs Learning Rate: Stability and Speed
Engineer training stability using batch size and LR scaling rules.
81
Regularization for Vision Nets
Dropout, label smoothing, stochastic depth.
82
Training Detectors
- Training Detectors 20 min
83
Annotation Strategy for Masks
Guidelines, QA, and edge-case policies.
84
Transfer for Detection/Segmentation
Backbone reuse and task-specific heads.
85
Training ViTs
- Training ViTs 19 min
86
Bias and Fairness
- Bias and Fairness 18 min
87
Real-Time Video Pipelines
- Real-Time Video Pipelines 18 min
88
Robustness Testing
- Robustness Testing 19 min
89
Monitoring CV Models
- Monitoring CV Models 18 min
90
Adversarial Robustness
- Adversarial Robustness 20 min
91
Data → Model → Output → Feedback Loop (Production Workflow)
End-to-end lifecycle: training, evaluation, deployment, monitoring, and iteration.
92
Edge Detection Deep Dive: Canny Tuning That Actually Works
Thresholds, hysteresis, and practical tuning workflow.
93
Preprocessing for Low-Light and Noisy Cameras
Denoise carefully and avoid destroying features.
94
Classical Pipeline: Features → Matching → Geometry → Decision
See the end-to-end classical workflow and where it still fits.
95
From Scratch vs Pretrained CNNs: A Cost/Benefit Guide
Know when training from scratch is worth it.
96
Designing Custom CNNs
- Designing Custom CNNs 18 min
97
Detection Evaluation: mAP
- Detection Evaluation: mAP 18 min
98
Segmentation Metrics
- Segmentation Metrics 18 min
99
Self-Supervised Pretraining
- Self-Supervised Pretraining 20 min
100
Transformers for Detection: DETR
Set prediction and reduced post-processing.
101
Privacy & Compliance
- Privacy & Compliance 17 min
102
Video Evaluation
- Video Evaluation 17 min
103
Error Analysis Workflow
- Error Analysis Workflow 17 min
104
CI/CD for Models
- CI/CD for Models 18 min
105
Privacy-Preserving Vision
- Privacy-Preserving Vision 18 min
106
Mini Project Blueprint: Ship Your First Image Classifier
A step-by-step blueprint that’s small but production-minded.
107
Frequency Domain Basics: FFT and Filtering
Understand low-pass/high-pass filtering and when it’s useful.
108
Production Preprocessing Checklist: Avoid Silent Accuracy Drops
A practical checklist for production pipelines.
109
Decision Guide: Classical vs Deep Learning in 2026
Choose the simplest method that meets your SLA.
110
Mini Project Checklist: Build a CNN Classifier End-to-End
Blueprint: loaders, transforms, training loop, eval, export.
111
Benchmarking Architectures for Production
Pick a backbone using latency/throughput/memory on target hardware.
112
Deploying Detectors in Streams
Queues, frame policies, batching, and edge deployment.
113
Real-Time Segmentation
- Real-Time Segmentation 18 min
114
Common Pitfalls
- Common Pitfalls 18 min
115
Backbone Selection in 2026
- Backbone Selection in 2026 15 min
116
System Design: Enrollment to Monitoring
End-to-end biometric product architecture.
117
Surveillance Analytics Case Study
Detection + tracking + events + alerts architecture.
118
Online Monitoring
- Online Monitoring 18 min
119
Edge Deployment
- Edge Deployment 19 min
120
Research Trends 2026
- Research Trends 2026 16 min
🎯 Interview Preparation
Top Python interview questions organized by difficulty
Nagmani Solanki
Digital Marketing
Edugators platform is the best place to learn live classes, and live projects by which you can understand easily and have excellent customer service.
Saurabh Arya
Full Stack Developer
It was a very good experience. Edugators and the instructor worked with us through the whole process to ensure we received the best training solution for our needs.
Praveen Madhukar
Web Design
I would definitely recommend taking courses from Edugators. The instructors are very knowledgeable, receptive to questions and willing to go out of the way to help you.
Need To Train Your Corporate Team ?
Customized Corporate Training Programs and Developing Skills For Project Success.
Get Newsletter
Subscibe to our newsletter and we will notify you about the newest updates on Edugators
