2. Using machine learning and computer vision technology, ViSenze can recommend visually similar items to online shoppers. Other companies claim to do this as well — what are the key advantages of ViSenze’s technology?
The space in which we operate, namely visual search and image recognition have been around for a while. It’s not a new space. Every company focuses on a couple areas in order to be better and sharper than their competitors. Similarly, for ViSenze, we’re not a general visual search company and, in every vertical that we focus on, we get down to the granularity and specificity of the domain itself.
We get down very deeply in understanding different visual attributes. For instance, in the fashion space, you need to understand the distinction between a mandarin collar and a standard collar. You also need to understand different types of sleeves: is it a half-shoulder sleeve? A cap sleeve? A quarter sleeve? We get down to that level of specificity in the very same way that a shopper would walk into a store and look for a specific dress knowing what she wants.
If we can at least equal that experience online by aggregating all these search attributes into an intelligent method of recommendations, we would have been very helpful to the online shopper.
3. How did you and your team come up with the idea behind ViSenze?
ViSenze was borne out of the efforts of computer vision scientists and real-world entrepreneurs. The funny thing was that when we were looking at this space, we realized that people were searching but not finding. Why? They were not describing what they were looking for efficiently or effectively with the keywords that merchants were using. They were speaking a different language from that of merchants’. We said to ourselves: if we could take away the product language taxonomy disconnect, and just simplify it using pixels, we would have solved a tremendous part of the problem that currently exists.
4. How do you see computer vision and machine learning changing the e-commerce and retail industry over the next decade?
Major retailers from all over the world, especially in the US, are already experimenting with machine learning. Right at the very top, the fundamental problem they are trying to solve is one of search.
The second problem they are trying to solve is that of recommendation. Machine learning has its application with NLP, price data optimization, site optimization and so on. In computer vision, it’s all about visual inputs. People are searching using images. So I see the combination of both machine learning and computer vision as fundamentally transforming the way that people are shopping, especially in visual merchandise.
All of us are attracted by products we see on the rack or the shelf. We are always attracted by the looks first. After looks, we base our decision on price, product availability, and so on. So, the way that these two disciplines are being combined fundamentally solves a major problem in search and recommendation.
5. What are the next set of projects that you are excited to take on? What new features are you looking to add to your platform?
ViSenze fundamentally has three technology stacks. One is visual search, the other is image recognition, and the final one is video. To speak to the video stack, very few companies are processing videos at scale using computer vision and deep machine learning techniques. ViSenze has patents backing our technology around the video space. If we can process videos in the same way we process images, there’s a lot more information and intelligence we can gain out of video. So that’s one of the key areas we will be doubling down on going forward.
With over a decade of AI experience, Digamma.ai’s team are your trusted machine learning consultants, partners, and engineers