L’Oreal has filed a patent for a method to determine directions in GAN latent space for controlling semantics in synthesized images. By utilizing auxiliary networks, the method allows for disentangled controls over GAN output, requiring minimal labeled data and enabling manipulation of latent code channels. GlobalData’s report on L’Oreal gives a 360-degree view of the company including its patenting strategy. Buy the report here.
Smarter leaders trust GlobalData
According to GlobalData’s company profile on L'Oreal, Nanoemulsion cosmetics was a key innovation area identified from patents. L'Oreal's grant share as of January 2024 was 61%. Grant share is based on the ratio of number of grants to total number of patents.
Controlling gan output semantics for image generation
The patent application (Publication Number: US20240037870A1) describes a method involving the generation of synthesized images using a generator with a latent code to manipulate target semantic attributes. The process includes discovering a meaningful direction at the latent code for the target attribute, defining a new code based on this direction, and outputting the synthesized image. The method also involves disentangling data directions, filtering out important dimensions from other semantic attributes, and optimizing the latent code to further manipulate the attribute. Additionally, the method allows for the manipulation of multiple semantic attributes, training other network models using the synthesized images, and receiving inputs to control the attributes in the generated images.
Furthermore, the patent application details the use of an auxiliary network with binary classifiers for each semantic attribute, providing meaningful directions for attribute control. The method includes steps for disentangling attribute controls, receiving user inputs to identify target attributes, and generating synthesized images for various applications such as e-commerce, recommendations, and augmented reality experiences. The patent also covers the manipulation of specific semantic attributes like facial features, poses, makeup effects, and cosmetic surgery effects. Overall, the method outlined in the patent application offers a comprehensive approach to generating synthesized images with controlled semantic attributes, enabling various applications in the fields of image manipulation and augmented reality.
To know more about GlobalData’s detailed insights on L'Oreal, buy the report here.
Premium Insights
From
The gold standard of business intelligence.
Blending expert knowledge with cutting-edge technology, GlobalData’s unrivalled proprietary data will enable you to decode what’s happening in your market. You can make better informed decisions and gain a future-proof advantage over your competitors.
GlobalData, the leading provider of industry intelligence, provided the underlying data, research, and analysis used to produce this article.
GlobalData Patent Analytics tracks bibliographic data, legal events data, point in time patent ownerships, and backward and forward citations from global patenting offices. Textual analysis and official patent classifications are used to group patents into key thematic areas and link them to specific companies across the world’s largest industries.
