jketreno/ketr.photos
1
0
Fork 0
Code Issues 1 Activity
ketr.photos/server
History
James Ketrenos 5d38cb4787 Show face boxes 
Signed-off-by: James Ketrenos <james_git@ketrenos.com>
2023-01-17 15:48:24 -08:00
..
db
Its working pretty well now
2023-01-10 16:29:11 -08:00
lib
Restructured scanner to be a little more maintainable
2023-01-12 12:52:11 -08:00
routes
Show face boxes
2023-01-17 15:48:24 -08:00
app.js
Ready to try!
2023-01-11 15:27:44 -08:00
console-line.js
Scanning and loading dirs
2018-01-15 20:31:13 -08:00
db-test.js
Added a few DB sanity tests for BLOB values between JS and Python
2023-01-16 17:06:55 -08:00
development.location
Docker react app loading.
2023-01-15 19:50:40 -08:00
face-recognizer.js
Remove all use of spread
2023-01-12 15:41:55 -08:00
face.js
Ready to try!
2023-01-11 15:27:44 -08:00
nginx.conf
Added nginx configurations
2023-01-15 17:22:01 -08:00
production.location
Docker react app loading.
2023-01-15 19:50:40 -08:00
README.md
Added server/README to document image flow
2022-12-31 16:28:13 -08:00
scanner.js
Processed all 30,000 photos without issue
2023-01-13 11:37:29 -08:00

README.md

Face recognition

  1. Check if FILE has been scanned. If so, skip.
  2. DETECT faces in FILE.
  3. EXTRACT rotated / normalized FACE into FACE-DATA/ID Q. How to map FILE::FACAE to FACE-DATA/ID
  4. CLUSTER / RECOGNIZE all FACE-DATA/ID Q. Does this require N:(N-1) comparisons where N is number of total FACES, or can it be N:C where C is the number of "SAME" cluster?