Extending Cloud Vision API by Ray Sakai -TMLS#3

Reducing dimensions / generating data by existing AI or: How I learned to stop worrying about CNN.

Speaker: Ray Sakai / Reactive Inc.

Transcript

1. Extending Cloud Vision API : or How I learned to stop worrying about CNN Ray Sakai, Reactive Inc
2. Motivation : Learning Deep Learning without Hardcore Mathematics
3. Step 1 : Hand Copying Sutra
4. huh?
5. Step 1 : Hand Copying Sutra
6. Google Cloud Vision API
7. Face Landmarks / Angles
8. OK, let me proof
9. Step 1 : Pick landmarks as inputs
10. Step 2 : Selfie a ton Then tag each picture either 0 or 1
11. Step 3: Get results from CVE Then JSON -> CSV with 0 or 1
12. Step 4: Training
13. DEMO
14. Future Steps : more data, better model
15. THANK YOU