byLOOK is image recognition software for smartphones that recognizes fashion and accessories and links you directly to the right shopping portals. This technology provides an easy-to-access, real-time solution for fashion portals, accelerating the sales process. Additional modules are being developed and will be available via byLOOK, making online shopping an even more enjoyable experience.
A couple of years ago, we discussed a project called byLook.
I had the idea to create a shopping app that recognizes fashion and accessories and automatically finds the stuff in a database, then linking to a shopping module.I hate shopping but seeing something in a magazine or at the street that you really like, recognizing it on a mobile and then buying it – this would be perfect for me! So I discussed this idea with our development team in Seattle. Jo and Jason – our main nerds – were excited. They exactly understood the benefit for women who hate malls but like the convenience of their smartphone… scan, click, buy…that’s the way it should be.
It turned out that the biggest challenge was to identify the clothing, or part of it, if you just point the camera to a person. The problem was much more involved than taking a picture and looking up that exact picture in a repository. We experimented with sophisticated feature recognition software as it is provided e.g. by MathematicaTM.
Here is an example: Assume you have me wearing these boots. You like the boots and want to buy the same. You take this picture:and you are interested in finding the boots!
We used commands like ImageKeypoints to find key features in an image. Then we extracted the main characteristics of those key points as small thumbnails. Here is how it looks (key points are the white dots).
We then compared these to “prototypes of shoes” trying to select the “class” of shoes (here “boots”) which then in turn provides a more detailed link to shopping for “brown boots”.But things got messy and we realized that Amazon with its huge inventory of things can do an easier job and indeed they did – we decided to not further pursue this experiment.
But now listen to this:
That’s exactly what happened….fast forward 3 years and Amazon’s “Fire Phone” emerged.
However, what I like about having disruptive ideas, discussing them and doing experiments or skunk work, even if it never comes to market: intelligence will never stop being beautiful. I like that vibrant community of people trying and having the drive to do it! How much I enjoy an afternoon sitting at Café Venetia on University Ave in Palo Alto and discussing new ideas… you hear them humming all around! And that’s what this is all about….!