The improvement in performance of the Internet permits today to broadcast over the Web films, video, live events (such as sports events, shows, conferences, etc.) and even TV programs and TV advertising spots with rapidly improving quality. Live@Web.Com aims to leverage upon this possibility and upon advanced technologies in the area of image processing and analysis, to create an innovative E-Commerce platform, which represents a real breakthrough versus current E-Commerce approaches and systems. LIVE@WEB.CM will allow companies to promote brands and products via videos broadcast over the Web, including live events. Live@Web.Com will allow consumers to perform on-line purchase by directly and real-time interacting with the running videos. Consumers can perform their searches and selections of products by clicking on brand-names, products samples, etc. while they appear in the running video. Once selections are made, purchases can be then finalised on-line by, e.g., comparing side-by-side products, getting more information on the selected products, ordering and paying, using more traditional E-Commerce techniques.
The project will develop an innovative E-Commerce platform, composed by 3 core modules:
Product Database. This module is a database including: images related to those products which are sold via Live@Web.Com and featuring them in different viewpoints; other product-related information (descriptions, details, prices, link with existing web-sites, etc.), typical of traditional E-Commerce systems. The links between the TV program and the advertising content have not to be defined explicitly in advance: product information will be created and maintained separately from the TV programs, and the links between the two will be dynamically created by the system through an image matching process.
Matching Engine. The Matching Engine includes the image processing algorithms, which are needed to extract objects from still and moving images and to perform searches in the product database by matching an input image with images in the database. Algorithms include edge detection, shape extraction and wavelets. The Matching Engine will be robust vs. image distortion, difference in size and proportions of searched objects and possible object misplacement. The system will be able to retrieve the selected product both in the case the consumer has selected a single frame in a video and in the case of the selection of an entire section of video.
E-Commerce Storefront. This module includes the functionality needed to broadcast TV programs in a Web page, accessible by the consumers via a normal Internet browser either via an Internet-connected TV-set or via a PC. It will also enable the consumer to interact with the TV program to select a single frame in the video or an entire section of it, and to store them for later replay or further examination, by placing them in a shopping tolley. It will also integrate with existing, state-of-the-art E-Commerce platforms, to offer more traditional functionality, such as navigating across product-related information; finalising on-line orders and secure on-line payments; etc. It will finally encapsulate the Human Computer Interface of the LIVE@WEB.COM system, through which the advanced capability offered by the Matching Engine will be available.
Funding SchemeCSC - Cost-sharing contracts
2132 JH Hoofddorp