The platform is developed as a web application and consists of two main modules that facilitate both the requesters and the contributors in designing and contributing to food knowledge crowdsourcing tasks. The user interface is designed to allow requesters a high degree of freedom to select contributors based on specific attributes, e.g., demographic profile, expertise level, engagement level, etc.
To handle a large number of food items, our task creation’s user interface allows requesters to flexibly select a subset of food items by specific attributes. Once created, tasks will be automatically shown to selected contributors in an intuitive user interface. Food companies looking to embark on an ongoing project to collect a large amount of public opinions about their products, e.g., flavor perception, visual appeal, can easily utilize our technology to set up their own online crowdsourcing platform and engage with their own panel of contributors.
Our technology is suitable for any organization looking to run their own food knowledge crowdsourcing platform. Note that the organization is required to provide their own food image database, requesters, and contributors. For example, a food company may set up a crowdsourcing platform using our technology, recruit a group of paid/volunteered contributors, and ask them to describe the flavors of the company’s products. Another example is to ask the contributors about the products’ visual appeal.
Furthermore, the technology can also be extended to set up a generic image annotation crowdsourcing tasks, for example, labeling other types of objects, as long as the tasks can be designed as a simple multiple-choice question.
Potential markets include companies in food and beverage (F&B) industry and food-science related organizations. As data and automation is becoming more important in the industries, many companies are looking to explore the use of crowdsourcing to enhance their business operations. Crowdsourcing platform helps bridge the gap between man-machine capabilities, allowing machine-learning software to perform various perception tasks normally performed by humans, e.g., recognizing products, analyzing product content.
The common feature provided by our system makes it easier and more flexible for organisations to run their own end-to-end food crowdsourcing project covering both the requester and contributor sides of the crowdsourcing market, via a platform out of the box. Once the system is properly setup, requesters and contributors can readily participate in the crowdsourcing efforts, e.g., creating tasks, assigning tasks to contributors, answering questions, by simply creating a user account through the system.