ET BureauBy Shilpa Elizabeth
CHENNAI: Indian consumers may soon have a reason to shop more online as fashion portals test an innovative product from a Chennaibased startup that makes it easy to find the perfect outfit online.
Mad Street Den, founded by neuroscientist Anand Chandrasekharan and product designer Ashwini Asokan, helps online retailers to display apparel that is exactly what customers are looking for, when they click and upload a picture of a preferred garment on the search bar of the website.
The technology empowers a site to recognise the cut, colour, texture and gender of the dress and help it to instantly bring back similar results using a mix of artificial intelligence and computer vision.
“The technology requires new user behaviour which is a very difficult thing to develop,” said Asokan who co-founded the venture about 18 months ago with both founders pooling in about $50,000.
That is one of the main reasons why they chose the ecommerce field to start with, because incorporating this technology into the online stores would demand only an extension of the existing user behaviour rather than creating a new behaviour.
The company, which launched the product two months ago, is already working with three ecommerce companies. Asokan declined to name them.
“Given that we are building a technology that is unique and differentiated, scaling too quickly without testing can be fatal,” she said. As of now, recommendations given by online stores are based only on brands. But the visual nature of the technology developed by the company has the potential to bring the customer much more varied results.
Over time, the site will get to know its users and will be able to build a history of personal preferences.
If a dress worn by a friend or displayed at a store is interesting, a user can click a picture of it and upload it on the search bar of the ecommerce site to find a match. Apart from such applications — which involve visual search and object recognition — the team has also developed other features that help detect emotion and decode expression.
“In expression and emotion detection, we are exploring options with gaming companies, ad agencies and TV and movie research companies, to know how people are reacting when something is going on the screen,” said Asokan.
At present, according to Asokan, such inputs are derived with the use of expensive hardware including high-end cameras. She claims her company has developed the capability to “bring all these features to the (cheapest) smartphone.”
Experts caution that the growth prospects for the company will depend on how robust the back-end technology support is.
“The question is how much reach an Indian company in this field will have outside and within India,” said Balram Raveendran, a professor at the Indian Institute of Technology, Madras, who specialises in artificial intelligence.
“The front-end seems pretty lightweight as it involves just clicking a picture and uploading it. But the back-end will essentially have to run on a server, so that’s where the heavy duty would come in,” he said.
CHENNAI: Indian consumers may soon have a reason to shop more online as fashion portals test an innovative product from a Chennaibased startup that makes it easy to find the perfect outfit online.
Mad Street Den, founded by neuroscientist Anand Chandrasekharan and product designer Ashwini Asokan, helps online retailers to display apparel that is exactly what customers are looking for, when they click and upload a picture of a preferred garment on the search bar of the website.
The technology empowers a site to recognise the cut, colour, texture and gender of the dress and help it to instantly bring back similar results using a mix of artificial intelligence and computer vision.
“The technology requires new user behaviour which is a very difficult thing to develop,” said Asokan who co-founded the venture about 18 months ago with both founders pooling in about $50,000.
That is one of the main reasons why they chose the ecommerce field to start with, because incorporating this technology into the online stores would demand only an extension of the existing user behaviour rather than creating a new behaviour.
The company, which launched the product two months ago, is already working with three ecommerce companies. Asokan declined to name them.
“Given that we are building a technology that is unique and differentiated, scaling too quickly without testing can be fatal,” she said. As of now, recommendations given by online stores are based only on brands. But the visual nature of the technology developed by the company has the potential to bring the customer much more varied results.
Over time, the site will get to know its users and will be able to build a history of personal preferences.
If a dress worn by a friend or displayed at a store is interesting, a user can click a picture of it and upload it on the search bar of the ecommerce site to find a match. Apart from such applications — which involve visual search and object recognition — the team has also developed other features that help detect emotion and decode expression.
“In expression and emotion detection, we are exploring options with gaming companies, ad agencies and TV and movie research companies, to know how people are reacting when something is going on the screen,” said Asokan.
At present, according to Asokan, such inputs are derived with the use of expensive hardware including high-end cameras. She claims her company has developed the capability to “bring all these features to the (cheapest) smartphone.”
Experts caution that the growth prospects for the company will depend on how robust the back-end technology support is.
“The question is how much reach an Indian company in this field will have outside and within India,” said Balram Raveendran, a professor at the Indian Institute of Technology, Madras, who specialises in artificial intelligence.
“The front-end seems pretty lightweight as it involves just clicking a picture and uploading it. But the back-end will essentially have to run on a server, so that’s where the heavy duty would come in,” he said.
Get Unlimited Access to The Economic Times
