COVID Detection - Deep Learning Solution on power 9 AI Box
summary
A VGG-16 transfer‑learning pipeline trains on roughly 11,000 chest x‑ray images (augmented from public pneumonia sets plus local hospital scans) and runs on a low‑cost Power9 AI Development Box, using PowerAI Vision for automated training and a simple web/mobile front end. The model reports about 95% classification accuracy after augmentation and pruning, serves real‑time inference on the device, and is packaged as a GUI web app for upload, classification and basic patient metadata.
AI-written summary · about
chapters (10)
- 0:00Introduction and objectives
- 2:15Problem definition and requirements
- 5:14Data collection and model training
- 6:28Transfer learning approach
- 8:13Data sources and model setup
- 10:03Model accuracy and results
- 12:18Web application and user interaction
- 13:03Future improvements and challenges
- 14:19Conclusion and next steps
- 19:10Q&A session
related talks
21mPowerAI Vision & Other Solutions on an AI Box
24mBreast cancer detection with IBM PowerAI Vision
29mCancer diagnostics enabled by machines: finding and classifying single tumor....
15mBrain Tumor Detection and Classification Using Deep Learning Models
01h 03mThe Ethical Dilemma of AI: Save the World or Forsake It?
55mDeep Learning in Medical Image Diagnostics
37mAI in medicine: where do we stand?
02mAI Tech's Smart City solution