ViSenze Blog

Visual search just got smarter: Introducing our Automatic Object Recognition feature

Posted by ViSenze 11-January-2016

With our newly launched Automatic Object Recognition feature, your users no longer have to crop an image query to focus on an item within the image, or manually select the product category in which to search.

After months of hard work, this is finally live.

How Automatic Object Recognition came about

As 2015 drew to a close, we questioned ourselves: what is the one feature that could push the envelope of visual search technology, and dramatically improve the user search experience? After speaking to customers, gathering business intelligence and pulling data from our dashboard, we reached a conclusion:  

To make visual search even better, faster and easier for end users, we need to fully automate every single step of the search process - from the instant an image query is uploaded, to the final generation of search results.

Letting our technology lead and take over the search process fully without a shred of human assistance was the next step. And we've managed to make that happen with Automatic Object Recognition.

How it works

This new feature - currently in beta version - has the ability to automatically detect and recognize the objects present in an image query before searching for other visually similar items and presenting results. In other words..forget about "fat finger moments" and the frustration that ensues when users find difficulty in cropping right against images with noisy backgrounds. No more of such manual work!

Allow our technology to take over. It knows which object in the image to focus on, and carries out the search process right away without users having to tell what it should do! By cutting out steps in this visual search process, the search experience becomes even more streamlined, intuitive, and easier.

automatic object recognition technology by ViSenze

The first of its kind in the visual technology space, the Automatic Object Recognition feature is able to detect various types of items, like tops, dresses, skirts, bottoms, shoes, bags, watches, Indian ethnic wear, jewelry, eyewear, and even furniture. 

Towards a future of automation and true intelligence

Besides further improving and fine-tuning the feature for specific cases and requirements, our team will continue to explore other product categories and jump on interesting project requests. Now, you can provide an extremely smooth user experience to your consumers. Full automation is your key to the future - a future of customer delight, conversion, and ultimately, loyalty. To try out Automatic Object Recognition on your database, simply contact us.  

For a seamless implementation of the feature for your desktop or mobile platform, follow these steps outlined here:

SDK instructions on python, Java, PHP, iOS, Android

 

  New product features

Comment