Computer Vision; Data Science; AI Applications by Domain

Computer Vision

What is Computer Vision?

Computer Vision enables machines to derive meaningful information from visual data (images, videos) and take actions or make recommendations based on that understanding.

Uses deep learning, primarily Convolutional Neural Networks (CNN).

Computer Vision Pipeline

Image Capture --> Preprocessing --> Feature Extraction --> Classification/Detection --> Action
  (Camera)      (Resize,Normalize)   (CNN layers)         (SVM, Softmax)          (Output label)

Key CV Tasks

Task	Description	Input	Output
Image Classification	Assign label to whole image	Image	Class label
Object Detection	Locate and classify objects	Image	Bounding boxes + labels
Semantic Segmentation	Classify each pixel	Image	Pixel-level map
Instance Segmentation	Segment each object instance	Image	Individual object masks
Pose Estimation	Detect body keypoints	Image/Video	Skeleton coordinates
Image Generation	Create new images	Text/noise	New image
Optical Flow	Track motion between frames	Video frames	Motion vectors

Domain	Application
Content Moderation	Filtering unsafe images/videos on social platforms
Facial Recognition	Phone unlock, security, Facebook DeepFace
Image Classification	Google Lens identifies flowers, animals, landmarks
Object Detection	Tesla/Waymo identify lampposts, pedestrians, traffic signals
Medical Imaging	X-ray/MRI/CT analysis for tumor detection
Agriculture	Drone-based crop disease detection
Retail	Amazon Go cashier-less shopping
Manufacturing	Automated defect detection on production lines
Augmented Reality	Snapchat filters, Pokemon Go
Document Processing	OCR, cheque reading, form scanning

Data Science

Definition

Data Science is an inter-disciplinary field that uses scientific methods, processes, algorithms and systems to extract knowledge and insights from structured and unstructured data.

Related to data mining, machine learning, big data.

Data Science Applications

Recommendations — personalized content
Price Prediction — real estate, stocks

Weather Forecasting — short and long-term

Data Science Workflow

Business Problem
     |
Data Collection (RDBMS, APIs, Web Scraping, Sensors)
     |
Data Cleaning & Preprocessing (missing values, outliers, normalization)
     |
Exploratory Data Analysis (EDA) — statistics, visualization
     |
Feature Engineering (create meaningful features)
     |
Model Building (ML/DL algorithms)
     |
Model Evaluation (accuracy, precision, recall, F1)
     |
Deployment (REST API, web app, dashboard)
     |
Monitoring & Maintenance

Key Data Science Libraries

Library	Purpose
NumPy	Numerical computation, arrays
Pandas	Data manipulation, DataFrames
Matplotlib / Seaborn	Data visualization
Scikit-learn	Traditional ML algorithms
TensorFlow	Deep learning (Google)
PyTorch	Deep learning (Facebook/Meta)
Keras	High-level deep learning API

AI Applications by Domain

Domain	Applications	Technologies
NLP	Speech recognition, translation, sentiment analysis	Transformers, BERT, GPT
Computer Vision	Facial recognition, object detection, medical imaging	CNN, YOLO, ResNet
Robotics	Manufacturing, healthcare, logistics, surgery	Computer vision + RL
Recommendation	E-commerce, streaming, social media	Collaborative filtering, DNN
Finance	Fraud detection, algorithmic trading, credit scoring	Anomaly detection, XGBoost
Healthcare	Disease diagnosis, drug discovery, medical imaging	AlphaFold, CNN
Chatbots	Customer support, personal assistance	NLP, Transformers
Gaming	Game AI, NPC behavior, opponent simulation	RL, MCTS
Smart Home	Device control, energy optimization	Edge AI, RL
Cybersecurity	Threat detection, intrusion detection, anomaly detection	Anomaly detection, ML
Agriculture	Crop monitoring, yield prediction, irrigation	CV + IoT
Transportation	Autonomous vehicles, route optimization	CV + RL
Education	Personalized learning, AI tutors	NLP + Recommendation
Legal	Contract analysis, case outcome prediction	NLP
Entertainment	Content generation, music composition	Generative AI