Electronic Business Innovation Service case: visual Product Search

Intermediary transaction SEO diagnosis Taobao guest Cloud host technology Hall

1, data mining and precision marketing services

2, visual product search (Image content identification search)

Representative Enterprise: Jiangsu View

Company website: www.1000look.net

Basic situation: Jiangsu View is a company providing visual search technology, founded in November 2011. In foreign countries this technology is also called icre (image content recognition Engine), that is, image recognition search. At present, one of the key products of Jiangsu View is to provide site image search technology services for E-commerce sites. When the user uploads the picture automatically identifies the image key area, extracts the text information of the commodity, automatically analyzes the color, style, shape, pattern and so on in the commodity picture, and so on, after the comprehensive judgment of all the information, selects the relevant products in the station database and the user's requirement, saves the user to find the commodity time, enhances the purchase conversion rate. Just a picture, you can let users in a large amount of merchandise quickly find the right baby, eliminating the choice of keyword description of the trouble.

Last month bukop.com in Beijing and the Jiangsu view of the entrepreneurial team had a brief communication, it is understood that its technical team is mainly abroad, in the field of image recognition with the world's leading level. Based on its cooperative analysis with North American electricity dealers, EPC (turnover/100 user access) has been greatly improved after using image similarity technology optimization. As a handful of domestic visual search engine Company, Jiangsu View Plan inherited the development path of the year, first as a technical provider of other sites, and then expand the common user market.

Visual search technology applied to E-commerce Web site's main function is two, one is to provide the search based on merchandise pictures, to help consumers achieve "retrace", to find the relevant products on the picture; the second is to provide a similar display of goods in the station and screening. For the style, design sense of non-standard goods especially applicable, such as clothing, jewelry. The ModCloth is a popular apparel online sales site in the United States, with the use of Site similarity optimization, EPC data rose nearly 5 times times.

Overseas similar companies:

incogna:2009 September, the integration of the country's most advanced product search function online shopping website Shopachu.com officially released, inside the use of visual search technology for users to find the required merchandise. Shopachu.com is the Canadian picture search engine Company Incogna's wholly owned subsidiary, Incogna's search speed is very fast, mainly based on color and shape similarity. (Judging from some public materials, the overseas technical team from Jiangsu View is probably from Incogna company)

Like.com: Time goes back to August 2010, when Google bought a company called Like.com at about 100 million dollars. The company has been chosen by Google because it has a technology that makes Google feel humiliated, image-recognition search. Google's original image search engine is to input text as a query criteria to select the picture, like.com is very cow, it also supports the input of pictures to query. If a user wants to find something similar through a picture, like.com uses something called a visual signature (visual Signature) to retrieve a similar result by comparing the visual signature of a user-supplied picture with another picture. Visual signature is actually a mathematical expression of a picture, which has a lot of characteristics, as long as two graphs have enough characteristics to be the same, like.com think they are similar. With this powerful technology, Like.com launched a series of small Web sites focused on fashion and E-commerce in 2009 to become a "visual search engine for goods" that allows users to browse websites to search for items of interest to them, or to find other items similar to their appearance. Depending on the search results, users can then select some of the features of the product pattern to further search for other items with similar styles, shapes, and colors. The weights of these three metrics can be set by the user to produce the search results that match their expectations. Also allow users to upload their own pictures, analysis of the site, returned with a similar appearance of goods. Prior to the acquisition of Like.com mainly through the Retail Site diversion volume to earn profits, according to CPS each order charge 10% Commission.

IQ engines: The company's any image recognition engine (any image recognition Engine) is a visual search engine that uses the exact classification algorithm and the VW outsourcing force to add tags to any picture. IQ engines intends to provide accurate, comprehensive picture intelligence services to shopping software, web Publishers, photo tagging software, and tools for people with visual impairments. The February 2010 IQ engines released its iphone based application Omoby, which provides a visual search for detailed descriptions of various items, user reviews, quotes, purchase links, and more.

Trend analysis: Visual image search by uploading images to search the same or similar picture information, known as the next generation of search engine core technology. The visual shopping search offers an innovative and exciting new online shopping experience. When you are walking on the road, or watching your favorite clothes on TV, if you can make a quick search, quickly find what you want, and realize the purchase, it is a perfect experience. The rapid development of E-commerce, the picture has become a commodity display, the best medium to facilitate the transaction, you can see the visual search method will become the main entrance of the network. Who has mastered the visual search, who will become the electronic commerce future user Traffic control.

This article is from bukop.com (author: He Yuqing)