From ae067b81ce1123a509526145d2e4acd2bf343cc2 Mon Sep 17 00:00:00 2001 From: Damien De Paoli Date: Mon, 6 Sep 2021 18:12:26 +1000 Subject: [PATCH] not worth fixing json* invocations, only 2, and the fix is either forcing a mutable pickle type or TypeDecorator -- both a bit complex than its worth --- TODO | 3 --- 1 file changed, 3 deletions(-) diff --git a/TODO b/TODO index 451bd3b..62ade08 100644 --- a/TODO +++ b/TODO @@ -5,9 +5,6 @@ START FORM SCRATCH for prod so all images have face_locn data - need to reconsider whether current distance algorithm gives best match - can I do better? - * could look to remove the hand fixing of json.loads of array data --> seems you can make your own datatype in the ORM, and it can do the conversion every time you use it - - https://stackoverflow.com/questions/28143557/sqlalchemy-convert-column-value-back-and-forth-between-internal-and-database-fo - * per file you could select an unknown face and add it as a ref img to an existing person, or make a new person and attach? * from menu, we could try to get smart/fancy... say find face with largest size, check it vs. other faces, if it matches more than say 10? we offer it up as a required ref img, then cut that face (with margin) out and use it is a new ref image / person - read that guys face matching / clustering / nearest neighbour examples, for a whole new AI capability