Visual Entanglement Analyzer (and Generator)

This web-based application is designed to find connections among geometric, floral, and calligraphic patterns across materials, times, and geographies of the pre-colonial Silk Roads—and beyond. While I created it primarily with these patterns in mind, it can be used with any set of images, including paintings or film stills.

Drawing on emerging Silk Roads scholarship, this tool traces relationships between patterns within and across images, revealing formal connections that cut across cultures and practices.

Built with Google AI APIs, this analyzer/generator accepts multiple images as input. Users can also optionally add annotations to each image to include information about form, significance, history, or any other context relevant to the analysis.

It then uses machine learning to segment the images and analyze each segment’s shape, content, composition, color, texture, and position within the overall image or pattern.

It then analyzes similarities between these “segment-features” (potential motifs) in terms of shape, composition, texture, and color, along with any contextual or historical connections included in the user’s annotations. Based on this analysis, it generates a graph where each node represents a segment-feature, and each link (or edge) indicates a similarity or connection between them. The strength of each connection is shown as a value between 0.0 and 1.0 next to each link.

The user can then select specific nodes and links to generate new images based on those chosen relationships and features.