Computer Vision & Edge AI
Teach machines to see. In 5 weeks, master convolutions, modern visual transformers (ViTs), and learn to optimize these heavy models for execution on drones, cameras, and mobile phones.
Mastered Technologies
You Will Build
Develop a real-time, edge-optimized defect detection system for a simulated factory line running inference at >30 frames per second on limited hardware.
The 5-Week Syllabus
An intense, week-by-week breakdown designed to push your limits.
Image Processing & CNN Foundations
The basics of pixels, filters, and convolutions.
Core Topics
- OpenCV Basics
- Convolutional Neural Networks
- Data Augmentation
Hands-on Lab
Build a custom CNN to classify manufacturing defects.
Modern Object Detection
Locating and classifying multiple objects in real-time.
Core Topics
- YOLOv8 Architecture
- Bounding Box Regression
- mAP Metrics
Hands-on Lab
Train a custom YOLO model to detect specific retail products on shelves.
Image Segmentation & Generators
Pixel-perfect classification and Generative Adversarial Networks.
Core Topics
- U-Net / Mask R-CNN
- Semantic Segmentation
- Intro to GANs
Hands-on Lab
Develop a medical imaging model that segments tumors from MRI scans.
Visual Transformers (ViT)
How attention mechanisms are taking over vision.
Core Topics
- Patch Embeddings
- Attention in Vision
- Hugging Face Diffusers
Hands-on Lab
Fine-tune a ViT for a highly specialized classification task.
Edge Deployment & TensorRT
Making models run fast on small hardware.
Core Topics
- Model Quantization
- ONNX Runtime
- NVIDIA TensorRT
Hands-on Lab
Optimize and deploy your object detection model to run at 30fps on a Raspberry Pi / Jetson Nano.
Expert Facilitator
Akiro spent 6 years deploying real-time object detection models to autonomous vehicles operating under heavy hardware constraints.
Student Perks
- Hardware discounts (NVIDIA Jetson)
- Kaggle Dataset Curation
- 1-on-1 Deployment Coaching