The time-travelling dhurrie

Museum of Art & Photography collaborates with Microsoft to launch an AI-powered, interactive digital platform that guides the viewer through textile history across regions and eras

A Persian kalamkari wall hanging that dates back to the late 17th Century or early 18th Century is my starting point. A closer look at the fabric reveals a royal personage seated in a pavilion surrounded by crocodiles and an elephant-like figure beside which stand seven disciples in courtly attire. The scene is perhaps set in a garden. Boasting the kalamkari technique of hand painting and block printing that involves resist dyeing, the fabric is a welcome mix of red, maroon and indigo, outlined with black.

A single click away, connected to the former by an artery, is a carpet adorned with flower motifs that originated between 1750 and 1850, possibly in Armenia or Dagestan. Aided by more informative arteries, a rich textile history reveals itself on screen. Thanks to AI (Artificial Intelligence) and machine learning powered by Microsoft, Interwoven, a vast collection of South Asian textiles that the Museum of Art & Photography (MAP) prides itself on, has just become globally accessible. 

The platform, which is fairly easy to navigate, has filters to explore connectors and differentiators between each piece of textile: patterns and motifs, geographies and cultures, date and eras, and the like. One also gets to devise a custom journey on selecting a single piece of art.

The still-growing collection of MAP has been in the works for one-and-a-half years.  “The reason we chose textiles is because it uncovers all these connections between regions and cultures. Take South Asian textiles for instance: our craftsmen have mastered some of the most complex weaving techniques and produced exquisite textiles that travelled over the seas because of the strong trade connections that existed. We have been in the world market for centuries,” says Kamini Sawhney, director, MAP.  Though Interwoven draws on the strong textile base in the museum’s collection, it also includes artefacts from institutes across the world, such as the V&A (London), MET (New York), Rietberg (Zürich), and the Royal Ontario Museum (Canada).

AI as a technology is clearly being adopted across multiple domains and sectors, Rohini Srivathsa, national technology officer at Microsoft India, quips. AI for Good is a Microsoft initiative that, over the next five years, will show significant commitment to use the technology to address some of the challenges as well as opportunities in the world. “AI for Cultural Heritage is one of five such programmes. And this collaboration is a marriage made in heaven,” she adds. 

“The thinking behind the experience when you go on the site, in terms of the journeys that people go through, comes from the MAP academy. Behind the scenes, there is quite a bit of technology that plays in — AI text analytics, cognitive search, custom vision, computer vision... There is a tremendous amount of tagging of data and developing of training models that happens in the background. But once you create that, the machine learning and AI engines are able to recognise some interesting connections between MAP and the partner museums,” says Rohini. These are connections that the human eye will perhaps fail to see. “It opens up abilities to create a very personalised experience,” she adds. When the right tools are placed within the hands of domain specialists and experts, the technology takes on the role of a multiplier. 

“There are two options for the viewer: the curated journey and the customised journey. The former is created by the educational research team at the MAP academy. For this, AI throws up connections between different global artefacts, and then the team researches and expands on them to create a journey. For example, in Leisure and Play, you would see motifs of hunting, reading or play. A colourful dhurrie from the 1940s depicts a hunting scene, leading you to an artefact from the Smithsonian Institute that shows the eagle dance from the 1830s.” Each time, the customised journey would be different (for the same viewer). Thus, an artist’s view would be different to that of a conservationist or a collector.