Computer Vision Bootcamp

What This Bootcamp Is About

Computer Vision is the branch of AI that has had the most industrial deployment — from manufacturing defect detection to medical imaging to autonomous vehicles. This bootcamp teaches students to build real CV pipelines: from raw images to annotated datasets to trained models to deployed APIs.

Curriculum

Day 1 — Image Processing Fundamentals (8 hours)

Module 1: How Machines See (1.5 hours)

• Pixel representations, color spaces (RGB, HSV, grayscale)
• Image as a tensor — shapes, channels, and batch dimensions

Module 2: OpenCV in Practice (2.5 hours)

• Reading, resizing, and color-space conversion
• Edge detection: Canny, Sobel
• Hands-on: Build a card/document scanner using contour detection

Module 3: Deep Dive into CNNs (4 hours)

• Convolution operation — filters, stride, padding
• Architecture evolution: LeNet → AlexNet → VGG → ResNet
• Transfer learning — why we don't train from scratch
• Hands-on: Fine-tune ResNet-18 on a custom 5-class image dataset (PyTorch)

Day 2 — Object Detection with YOLO (8 hours)

Module 4: Object Detection Fundamentals (2 hours)

• Detection vs. classification vs. segmentation
• Anchor boxes, IoU, non-max suppression
• mAP — the primary evaluation metric

Module 5: Dataset Preparation with Roboflow (2.5 hours)

• Image annotation workflow (bounding boxes)
• Dataset augmentation strategies
• Hands-on: Annotate a small custom dataset

Module 6: Training YOLOv8 (3.5 hours)

• YOLOv8 training configuration
• Evaluation: mAP, precision-recall curves, confusion matrix
• Hands-on: Full training run on the custom dataset

Day 3 — Real-Time Inference and Deployment (8 hours)

Module 7: Running Models on Video (3 hours)

• Frame-by-frame inference on video files
• Webcam inference with OpenCV
• Hands-on: Real-time helmet/safety detection demo

Module 8: Deploying a CV API (3 hours)

• FastAPI basics — endpoints, file uploads, responses
• Wrapping a YOLO model in a REST API
• Dockerizing the API
• Hands-on: Ship a working /detect endpoint

Module 9: Capstone Project (2 hours)

• Teams pick one: pothole detection, attendance system, defect detection, PPE compliance checker

AICTE Activity Points

Eligible for AICTE activity points under the Technical Activities category.

Inquiry

Want to run this bootcamp at your college? Request a proposal →