Azure Computer Vision
Definition
Azure Computer Vision is Microsoft's AI service that analyzes images and extracts valuable information from visual content for various applications.
Use Cases
- Microsoft: Accessibility features that describe images and read text for people with low vision — Microsoft’s Seeing AI mobile app uses cloud-based computer vision capabilities (image description, OCR, and object recognition) to interpret scenes and read printed text aloud. (Improves accessibility by helping users understand surroundings and printed materials more independently.)
- Walmart: Retail shelf and inventory visibility from images — Walmart has used computer vision in stores to interpret images/video for shelf conditions and product availability signals, integrating outputs into store operations workflows. (Supports faster detection of out-of-stocks and shelf issues, improving on-shelf availability and operational efficiency.)
- Uber: Document and identity verification from photos — Uber has used computer vision and OCR techniques to extract information from uploaded documents (e.g., IDs) and to support verification flows in its onboarding and compliance processes. (Helps streamline onboarding and compliance checks by reducing manual review effort and speeding up verification.)
Provider Equivalents
- AWS: Amazon Rekognition
- Azure: Azure AI Vision (Image Analysis) / Computer Vision
- GCP: Google Cloud Vision API
- OCI: OCI AI Vision
Frequently Asked Questions
- What's the difference between Azure Computer Vision and Azure Custom Vision?
- Azure Computer Vision (Azure AI Vision Image Analysis) provides prebuilt models for common tasks like image captions, object detection, OCR, and content moderation. Azure Custom Vision is for training your own image classifier or object detector using your labeled images (for example, recognizing your specific product SKUs or defects). Use Computer Vision for general understanding; use Custom Vision when you need domain-specific recognition.
- When should I use Azure Computer Vision?
- Use it when you need to extract information from images without building and training your own model. Common scenarios include OCR for receipts/forms, auto-tagging images in a DAM system, generating captions/alt text for accessibility, detecting objects in photos for search, and flagging potentially unsafe content. If you need to recognize very specific items unique to your business, consider Azure Custom Vision instead.
- How much does Azure Computer Vision cost?
- Pricing is usage-based and typically depends on the number of transactions (images analyzed) and which features you call (for example, OCR vs. image analysis vs. more advanced options). Costs can also vary by region and pricing tier. To estimate accurately, identify the APIs you’ll use, expected monthly call volume, and whether you need higher throughput; then check the Azure AI Vision pricing page for the current per-transaction rates.
Category: ai-ml
Difficulty: intermediate
Related Terms
See Also