Multimodal Conversational Interfaces with GPT and Vision AI
summary
Demos show GPT‑4 Turbo with vision on Azure AI producing concrete outputs: generated six‑sentence rental listings from an uploaded photo, flood detection with actionable advice, OCR and receipt JSON improved by combining with the Florence model, graph images turned into runnable Python code, screw defect identification and a video anomaly detector that boils footage into 17 key frames and flags a fire, plus searchable product matches via Azure AI Search and extracted license plates.
AI-written summary · about
chapters (13)
- 0:00Multimodal AI with GPT-4 Turbo
- 2:00Vacation rental descriptions from images
- 5:13Flood reasoning from images
- 7:00Handwritten text extraction
- 9:11Python code from graph images
- 9:54Broken screws detection
- 11:00GPT-4 Turbo with Vision & Florence
- 12:50Drone footage analysis
- 15:17Car damage reports from videos
- 17:25Receipts processing with OCR
- 19:14Azure AI with custom images
- 23:56Seeing AI app overview
- 24:57Closing remarks & Q&A
related talks
39mNew multimodal vision AI models and their practical applications
46mMultimodel Multimodal and Multiagent innovation with Azure AI
51mThe Next Frontier: Azure OpenAI Models and the Arrival of GPT-4
48mGetting started with generative AI using Azure OpenAI Service
39mDisrupt and Grow with Microsoft Azure AI
24mInfuse Generative AI in your apps using
01h 10mSolutions prototyping using Azure AI Studio
16mUnlocking AI Innovation: A Conversation with Eric Boyd and Asha Sharma