This report explains what has been built before the first deployment: not a final recognizer of living structure, but an App Engine-compatible instrument for seeing where such a recognizer could become precise.
The first change from the earlier proposal is practical: the system is designed as a three-tier Google App Engine app. The browser handles upload and reports. Flask performs analysis. Datastore stores the compressed image, metrics, SVG diagnostics, and report.
The analyzer uses ordinary image operations: grayscale conversion, thresholding, connected components, edge maps, bounding boxes, centroids, and simple shape scores. Every number has a visible diagram attached to it.
The first metrics correspond to relatively operational properties: centers, levels of scale, positive space, boundaries, roughness, and local symmetries.
This version is a measuring instrument, not a finished theory. The wholeness score is deliberately labeled “initial.” Its purpose is to make disagreements visible. When it calls a dead space positive, or misses a living space between branches, that failure becomes data for the next phase.
Upload, downscale, analyze, store, render reports, and inspect the first center/space diagrams.
Human pairwise comparison, hand-corrected center diagrams, image categories, calibration sets, and better positive-space geometry.