Naver announced on the 28th that it has applied a multimodal artificial intelligence (AI) model to ‘Smart Lens’, its deep learning-based image search service.
Multimodal AI is a core technology for the next-generation search environment. It is an AI model that learns complex information such as text and images simultaneously to quickly and effectively find the results the user wants.
Unlike existing search environments where information had to be found only through a single means such as text, image, or voice, multimodal AI allows sophisticated customized searches by inputting complex information, such as combining images and text.
Last March, Naver built ‘OmniSearch’, a multimodal AI that learns big data in services such as blogs, cafes, shopping, and news, and matches keywords entered by users with a high level of search results, and applied it to services. It allows users to directly access the latest AI technology.
First, the model was applied to smart lenses. Smart Lens is a service that allows specialized searches on various topics, such as shopping, text recognition, and wine labels, when the user captures the image they want to search for. It is provided through mobile Green Dot.
In smart lenses equipped with multimodal AI, the ‘+Add search term’ function allows you to find detailed information by entering additional text after taking an image. For example, after searching for an image of a specific shoe, you can narrow down the product line you are looking for by adding additional search words according to the user’s preference, such as color, design, and material, in text.
Currently, it is available starting from the ‘sneakers’ category, and the scope of application will be sequentially expanded to various categories in the future. In addition, we aim to establish a differentiated search environment by introducing advanced omnisearch to all Naver searches, including shopping searches, within the year.
The company predicts that it will be possible to provide a search service specialized in the shopping field when OmniSearch, which is built on massive shopping-related big data, is applied.
Kang Yu-hoon, responsible leader of Naver Search’s independent corporation (CIC), said, “We plan to continue to lead the everydayization of AI so that users can experience Naver’s AI technology in real life through various search services.”
Source: Pangyo Techno Valley Official Newsroom
→ Go to ‘Asia Innovation Hub Pangyo Techno Valley 2022’ news