GAIKSemantic Dental Video Search
Intelligent search across dental instructional video libraries
Semantic Dental Video Search
Find the exact moment in a dental lecture where a topic is discussed — just by typing a question. The system searches across an entire video library and returns timestamped results, so you can jump directly to the relevant segment.
How It Works
The system processes each video through three stages:
- Subtitles are split into segments — Each video's subtitles are grouped into ~30–60 second chunks
- Each segment gets a numerical fingerprint — An AI model converts the text into a vector embedding that captures its meaning, not just keywords
- Search combines meaning and keywords — When you type a query, the system finds segments that are both semantically similar and contain matching terms
Search Methods
The system supports three search modes:
- Semantic search — Finds results by meaning. Searching for "tooth replacement" also finds segments about "dental implants" or "prosthetics", even if those exact words aren't used
- Keyword search — Finds results by matching specific terms in the text using full-text search
- Hybrid search (default) — Combines both methods using Reciprocal Rank Fusion (RRF). Results that rank high in both semantic and keyword search appear first. This is the most reliable mode for everyday use
Finnish Language Support
Finnish is an agglutinative language — words change form based on grammar (e.g. "xylitol" becomes "xylitolin", "xylitolia", "xylitolista"). Standard keyword search would miss these forms.
The system uses prefix matching: the search term "xylitol" automatically matches all Finnish inflections starting with "xylitol-". Combined with trigram similarity as a fallback, the system handles typos and word variations gracefully.
Precise Timestamp Seeking
Each 30–60 second segment is backed by individual subtitle lines (2–5 seconds each). When a search result is selected, the system resolves the exact subtitle line within the segment, achieving ~3–5 second precision for video seeking.
Getting Started
The database schema and search functions are provided as ready-to-use SQL migration scripts:
- Start a PostgreSQL database with the pgvector extension
- Run the 5 SQL migration files in order to create the schema, indexes, and search functions
- Insert video segments with embeddings from your AI model
- Call the
hybrid_search()function to search
SQL scripts and setup guide:
implementation_layer/src/gaik/software_components/RAG/pg_vector_store/sql/