श्रीModi ShatVidhiTransliteration workspace
Python pipeline OpenCV · local processing

A web interface for the manuscript-analysis pipeline

From Modi manuscript
to meaningful reading.

Upload an archaic Modi document and inspect every processing decision — from paper cleanup and line localisation to complete glyph candidates for transliteration.

01Acquire→02Prepare→03Locate lines→04Segment glyphs→05Recognise→

Ready for a document

Your analysis workspace

1–4Phase
VisualOutput
LocalMode
श्री

The document journey begins here.

Upload a scan from the control panel. Its real intermediate images will replace this panel after the Python pipeline completes.

01
Acquire

Upload a colour or grayscale scan. The server decodes it with OpenCV without altering your original file.

02
Prepare

Bilateral denoising, adaptive thresholding, component filtering, and HPP deskewing create a clean manuscript surface.

03
Locate lines

Autocorrelation estimates spacing; an anisotropic Gaussian merge finds text-line regions across the page.

04
Segment glyphs

Zonal analysis protects matras, then ligature, cluster, and kana rules form complete Modi akshara candidates.