Difference between revisions of "Sketchy recognition/2020"
Jump to navigation
Jump to search
(→Sketchy Recognition 2020 (Notes)) |
(→Working article) |
||
| Line 38: | Line 38: | ||
=== Working article === | === Working article === | ||
The working text / structure of the online publication is here: [[Sketchy_recognition_2020]] | The working text / structure of the online publication is here: [[Sketchy_recognition_2020]] | ||
| + | |||
| + | and NOW | ||
| + | |||
| + | https://diversions.constantvzw.org/vandalist/ | ||
=== Selection of drawings === | === Selection of drawings === | ||
Revision as of 15:10, 29 May 2020
Contents
Sketchy Recognition 2020 (Notes)
For the online installation of Sketchy Recognition, we plan a multiform online publication that (1) documents the experiences of the pianofabriek installation, (2) provides addition context in the form of text, software probes, and video, and (3) an invitation to participation through downloadable/printable PDF coloring book (photos of complete drawings can be uploaded to the site / archive), and an online discussion / collective drawing session.
Materials we have
- Cherchez l'erreur (sketch)
- Detourned carmentis -- pianofabriek interface + materials
- The animation cells -- photos taken in the installation as people were drawing (todo: are these complete?)
- Previous notes including links to first selection process of contours
- Saskia's story of the mislabelled musical instrument (todo: link to transcription)
Topics to address
- The non-neutrality of the frame, questioning the hygenic pseudo-neutral space of the carmentis / classical archival space of the museum.
- The non-neutrality / culurally-b(i)ased classifications that comes from the quickdraw data set
- The mismatch of how things are (mis) classified...
Links/Sources/Inspirations/Related work
- http://plottingd.at/a/ -- Online publication about datasets (COCO) -- Ruben van de Ven + Cristina Cochior
- https://quickdraw.withgoogle.com/
- In browser sketch recognition https://observablehq.com/@xaris/doodle-recognition-with-tensorflow-js
- https://cabreraalex.github.io/interactive-classification/
- https://github.com/tensorflow/tfjs/blob/master/GALLERY.md
- https://donottouchyourface.com/
Probes
- Contour drawing (todo: pick a exemplary form /object)
- Sketch recognition (version like the "original" project)
- Combing sketch recognition with contour drawing...
- Sketch Recognition + Contour + Rotation (demonstrate how recognition changes as an object is rotated)
Possible outcomes / forms
- Online publication with text, probes, video
- PDF / Extended Comic book
- Workshop / Online discussion / Collective drawing session
Working article
The working text / structure of the online publication is here: Sketchy_recognition_2020
and NOW
https://diversions.constantvzw.org/vandalist/
