Intense competition between Facebook and the Russian social network, Odnoklassniki, has given the visual recognition technology start-up, Kuznech, a chance to firmly establish itself on the Russian market and to enter the U.S. Market.
The Russian-American startup, Kuznech, was founded in 2010, and since then the company has won 10 international IT awards. In 2014, its revenues reached $700,000.
Kuznech’s success story began in 2012 when Facebook bought the face detection technology provider, Face.com. The latter then recalled its license from the Russian social network, Odnoklassniki, a popular competitor to Facebook on the local market. Kuznech, which is based in Boston and San Francisco, moved quickly to occupy that niche.
Technology inspired by insects
The company’s name derives from the Russian word for “grasshopper,” and the technology it uses truly has much in common with the green insect. Grasshoppers have multiple eyes and they see the world as a picture consisting of numerous dots. “Our technology is powered by neural networks that represent simplified models of the nervous system of living organisms,” said company co-founder, Michael Pogrebnyak. “In other words, the principles that our technology is based on are very similar to the processes that take place during visual recognition in the human brain: learning, generalization, abstraction.”
The idea behind the technology belongs to one of the company’s co-founders, Alexander Valencia Campo. He was developing computer games and constantly came up against the problem of searching for similar images of illustrations. Initial investment in the project was $500,000, with a further $750,000 invested in 2011-2012 by the Skolkovo Foundation.
Kuznech currently offers eight products, and the main ones are visual search and recognition for images and videos, as well as mobile recognition. The company holds four patents; one in Russia and three in the U.S.
Future in medicine
Kuznech’s main clients are news agencies, online stores, and social networks. The company’s development center is based in St Petersburg, and the big names on the Kuznech’s client list include Mail.ru Group (Russia), and PartsTown (USA). There are different ways to monetize the technology for various products. Annual non-exclusive licenses are the main one.
On the world market, especially in North America, many competitors offer similar technology, such as Idee. According to Pogrebnyak, however, the advantage of Kuznech’s technology is that it’s practically universal and can be built into different applications.
Pogrebnyak is convinced that this technology has huge potential in analyzing medical images. “We are trying to detect melanomas and other skin growths, and we plan to set up a melanoma cell detector for doctors,” added Pogrebnyak. “However, things have been rather slow as investors like the idea but are in no hurry to put money into it.”
Konstantin Vinogradov, an analyst with Runa Capital venture fund, believes that one of the most crucial challenges for startup companies in this field is lack of scalable, profitable and unoccupied market niches. “Kuznech can capture a part of the international market if it focuses on a few cases that its competitors haven’t, but it will be difficult because there are already a lot of players on the U.S. market and local players from Southeast Asia,” Vinogradov said.
You can try Kuznech’s Face Detection and Face Recognition demo here: http://facedetection.kuznech.com/
Read article оn rbth.com
By Paul Melcher @Kaptur.co
Image recognition, visual search and content classification have been around for a while with various degree of success. With mobile shopping exploding, as well as myriads of photo/video based platforms, it is now at the core of almost any online experience. While Google, Yahoo, Facebook and Microsoft throw millions in research for their own gain, independent companies offer powerful solutions to the rest of the market. 4 years old Russia-based Kuznech is one of them. We spoke with co-founder and CEO Michael Pogrebnyak to learn more:
— Explain Kuznech. What does it solve?
“Kuznech” is a shortened version of a Russian word “kuznechik”, which means grasshopper. The vision of grasshoppers is very different from that of human beings and many other creatures. Grasshoppers have 5 eyes in total: three small eyes on the top of the head and two large compound eyes. Each compound eye is made up of thousands of very tiny eyes called ommatidia. These miniature eyes take in small portions of light from the full image that a grasshopper is observing and build a big picture out of these “data”. We thought that a grasshopper could be a great symbol for a company working on image, video and object recognition.
— What kind of companies use your technology?
Mainly, these are companies that manage large databases of visual information (images and videos): dating sites, social networks, video and image hosting platforms. For all these types of companies, we provide content moderation services, like nudity detection (checking whether the UGC contains adult content or not), video and image search, avatar moderation. In this field, we work with Russian’s largest social networks Odnoklassniki and VKontakte (Mail.ru Group).
Then, it’s e‐commerce market, where we present our mobile recognition solution: “See – Snap – Purchase”. For example, a US company PartsTown (they sell genuine OEM replacement parts for restaurants and commercial kitchens) and Russian HSC (Helicopter Service Company) use our recognition technology in their mobile applications.
And the third use case is logo recognition in images and videos. Here we look towards sports sponsorship market and work with marketing managers of sports clubs.
Sports sponsorship is used by companies to enhance corporate image, increase brand awareness, and win global recognition with high-value consumer. According to IEG, sport sponsorships will be worth $40B worldwide in 2015. While the sponsorship industry continues to expand, robust measurement activities often lag behind. Thus, feedback on how frequently the logo is actually displayed within the sports video and/or social media images is truly valuable for the advertising industry. Advanced visual search technologies enable you to automatically detect and recognize logos in the high-motion setting of sports videos and create a report containing various kinds of statistics like: frequency of recognized logos, their visibility time and their locations in the video, – which, in turn, allows for advertisement efficiency verification.
Over the last years, social media monitoring has become a primary form of business intelligence. With special services you can track your brand’s mentions – respond to critics, engage in conversation, or simply learn about your customer’s interests and trends. Getting your brand present on different social platforms allows you to connect with your current and future customers from all walks of life.
In the year 2015, 66% of marketers agree that social media marketing is core to their business. And nearly double the number of marketers categorize social as a primary revenue source compared to 2014.
M-commerce is a growing trend within the tech industry. Shortly said, it’s the niche of E-commerce for your mobile telephone, which has skyrocketed since iOS and the new Android operating system. Maria Zhulkova, BizDev @Kuznech, talks about a high-performance visual image and video search technology.
Learn more on intoconnection.com
Brands are spending large amounts of money on sponsorships, in particular in sports, which are seen as a unique way of engaging emotionally with fans. Ideally the brand will be featured prominently in an image of a star player scoring a key goal for the home side and reap the benefits of being connected to a moment of collective glory. Anecdotally brands get “a lot” of exposure for their sponsorships of teams and athletes via images shared on social media, but up till now, no one has been able to quantify this valuable audience.
Luckily for brands, the convergence of existing computer vision technology and the recent advances in machine learning are changing the game. Large-scale analysis of social media images to identify brand logos and gather useful information about audience and engagement is now emerging as a credible approach to earned media measurement, especially for sport sponsorship. It is now possible to look inside the image to detect faces, objects and brand logos at a scale, speed and accuracy that was impossible a few years ago. These new approaches reveal huge audiences and high levels of engagement that were previously invisible.
The new services available from companies like Kuznech, combined with inexpensive cloud computing puts brand-spotting within reach for brand managers. Kuznech can do brand, logo, face recognition on images, as well as on videos.
Widespread adoption of these new technologies will change the way earned media is measured. Brands will have new ways to measure the ROI of their sponsorship dollars and engage directly with the audience that is already emotionally connected to their brand. Brands, agencies, teams and events that grasp these tools now will lead in the future.
Parts Town, a market leader in restaurant equipment parts distribution, launched their latest innovation: the entirely redone Parts Town app with revolutionary PartsMATCH parts identification technology and a dramatically improved in-app shopping experience.
PartsMATCH 1.0 utilizes cutting edge image recognition algorithms to compare customer-submitted pictures of a part against hundreds of thousands of PartSPIN images to find a match. While the technology is in its infancy and still evolving, Parts Town is committed to pushing the envelope with unique customer solutions.
The image recognition engine was developed by Kuznech Inc, a company that creates object recognition technologies for m-commerce. “Image recognition is being used in fields like national security, social networking, and retail. As we’ve done with other industry innovations such as PartSPIN, we applied an external technology to solve a parts-related problem”, said Mike O’Shea, Parts Town’s Director of Ecommerce. “This feature, developed by Kuznech, reinforces our commitment to drive the industry forward, and we will continue to refine and improve it over the coming months”.
PartsMATCH Assistant applies a human touch to the same parts identification challenge. Customers can submit parts identification requests right from the app, including pictures of the part, then receive push notifications direct to their device as soon as a PartsMATCH specialist has located their matching part. “At Parts Town we are committed to helping each of our customers find the right OEM part for their equipment. Pictures are often extremely helpful in the parts identification process, and PartsMATCH Assistant combines great technology and great people to get answers to customers faster than ever before”, said Steve Snower, President of Parts Town.
Significant shopping and speed improvements were also incorporated into the new app. “Technicians and operators need access to parts technology on the go now more than ever, and the all-new Parts Town app is an indispensable tool for our increasingly mobile customers”, added Snower. Initially on iOS only, an Android version will be available soon.
In accordance with London eCommerce Expo 2015:
- Mobile is critical to retail in the coming years. You need to give people a real reason to engage – games and fun stuff (snapping a product is a part of your gamification strategy). Add sufficient value that people keep coming back to you.
- Only 4% of budgets go into mobile but it’s where 20% of consumers time is.
- Mobile is disrupting commerce. Physical shop real estate is shrinking, so e-commerce is more important than ever.
Be smart, use Mobile Recognition
When we here at Kuznech think of all the benefits of being in our business, we quickly think of our relationships with great people like you.
We would like to acknowledge our partners, customers, investors, journalists, website visitors, and all you guys who have helped us shape our business and give us the chance to do what we enjoy.
Thanks for a great year, and we wish you all the best as you embark on 2015.
Michael Pogrebnyak, CEO, and your Kuznech Team
There are 100 photos of EU, US, and Russian celebrities in the demosystem, and you can upload your own images to the system to see how face detection and face recognition algorithms work. Here are our results.
We welcome comments and suggestions and are open to any and all feedback. Please feel free to share your opinion here.