Inter-Omnia-Glypha · OCR & Computer Vision
"The void between
characters
is computable."
Reversed binarization transforms voids into data —
steganographic · AI-native
Designed to read from downsampled images — potentially ~4× fewer pixels (½ resolution per axis); benchmark in progress.
Interval objects (Derrida p. 11 scan)
1,675
CER (synthetic DejaVuSans, 176 lines, Jul 2026)
1.7%
How It Works
Derrida inspired visualisation pipeline
Six transformation stages from raw image to AI-native glyph graph
Original
Reversed
Skeleton
CA
Canny
Guided
Live Demo
Upload an image.
Get OCR results in real time.
The Entreacte API processes your image through the full reversed-binarization pipeline and returns structured glyph data with bounding boxes, entropy scores, and steganographic metadata.
▶ Open DemoDrop image here
PNG, JPEG, TIFF, PDF
Presented at IEEE ICFSP 2026, Toulouse — 10–12 June 2026. Paper published. Contact us for collaboration →