6,700
edits
Changes
Jump to navigation
Jump to search
Line 63:
Line 63: − + Line 73:
Line 73: − +
Nutrition facts table data extraction (view source)
Revision as of 15:05, 10 July 2020
, 15:05, 10 July 2020→Planning
* Test of an end-to-end model, through an API as a starter: Form Recognizer or Google Document AI API. If results are promising, we develop in-house end-to-end model, otherwise a layout model.
* Test of an end-to-end model, through an API as a starter: Form Recognizer or Google Document AI API. If results are promising, we develop in-house end-to-end model, otherwise a layout model.
* Thourough litterature review
* Thorough literature review
* Architecture choice
* Architecture choice
* Annotation campaign
* Annotation campaign
Yichen: test of image preprocessing before Form Recognizer (layout model), test "manual" approaches.
Yichen: test of image preprocessing before Form Recognizer (layout model), test "manual" approaches.
Raphaël: Litterature review (to be continued)
Raphaël: Literature review (to be continued)
= Resources =
= Resources =