👁️ Part 1: How Machines "See"
The goal is interpreting pixels, not just recording them.
- **Template Matching:** Simple but rigid; only works with exact matches in size/style.
- **Feature Extraction:** Smarter approach; identifies shapes (lines, curves) and cleans data using image processing (edge enhancement, smoothing).
💬 Part 2: How Machines "Talk"
Human language is ambiguous and context-heavy, unlike strict computer code.
Three Levels of Analysis:
- **Syntactic:** Grammar check (identifying subjects/objects).
- **Semantic:** Meaning check (recognizing the same concept in different sentences).
- **Contextual:** Vibe check (understanding intent and external context).
📊 Bonus Round: Handling Big Text Data
- **Information Retrieval:** Smart searching beyond keywords (the "Semantic Web" idea).
- **Information Extraction:** Pulling specific facts to fill **frames** (templates) or build **semantic nets** (linked data).