How to Build a Cloud Native Image Recognition Solution
Talkat KubeCon + CloudNativeCon North America 2020 Virtual · on Wednesday, 2020-11-18· at 16:55· Intrado Virtual Event Platform,·
summary
A cloud-native pipeline that combines OCI API Gateway, Fn Project functions, Object Storage, OKE and Google AutoML OCR/NLP exposes a single API that classifies documents, extracts fields without pre-defined markers, and returns a quality gate 0.7 confidence score; examples show one document at 0.7 and another at 0.53. Field extraction uses a rastering position math approach to group OCR boxes, supports serverless and Kubernetes deployment, and cuts manual review time by about 90% in a Mexican bank pilot.
AI-written summary · about
chapters (13)
- 0:00Introduction
- 2:09The Use Case
- 9:08Oracle Cloud Services
- 10:27API Management
- 13:05Oracle Functions
- 14:52Oracle Container Engine
- 19:36Connecting the Dots
- 20:24The 7 Steps
- 21:23Components from Oracle
- 22:57Architecture
- 23:45Demo
- 26:18API Result
- 32:50Conclusion
related talks
31mIntelligent Documents Management System
15mIP Detection at Scale: CI/CD Meets Cloud-Native Enforcement
29mBetter Understand Your Documents with AI
29mHow to Effortlessly Automate File Recognition with AutoML
27mCloud Native Artificial Intelligence - Top to Bottom
36mCloud AI Services: What they are and how to use them
35mCloud-Native MLOps Platform for Python- and ML Engineers
51mCloud AI: How to Get Started Injecting AI Into Your Applications