Toggle3D.ai announced that the Company has developed a new AI photo segmentation algorithm for the Toggle3D web application. This cutting-edge AI-algo is poised to redefine the boundaries of image analysis and 3D texturing, unlocking a multitude of opportunities by attracting new customers. The new feature builds upon Toggle3D.ai's continuous development by adding efficiency and speed to the 3D texturing creative workflow by using AI-powered tools to save time and money for individual users and entire design teams. This feature will be just one of many in the suite of AI-powered tools being released in the platform in the coming months for Platform Pro users only.

This supports Toggle3D.ai's SaaS business model, as Platform Pro is a paid subscription available to users for $29/month. All the AI features that are being developed for the platform are specifically for Platform Pro users, and not available on the free version, enticing users to sign up. When users upload a 3D model and 2D reference image, the AI segments the 2D reference into parts, which users can select directly in the platform.

The technology then takes a patch from the part and searches that texture patch against the pre-existing material library, automating the matching process from the 2D photo to the 3D material. Photo segmentation, the process of partitioning an image into distinct, meaningful regions, has traditionally been a labor-intensive and time-consuming task, requiring manual intervention. However, Toggle3D.ai's new algorithm represents a significant leap forward by automating this process with unparalleled accuracy and efficiency, real-time performance, and seamless integration.

There are 3 key steps in the AI photo segmentation process: Segmentation: The different regions in the input image are automatically segmented using transformer-based AI model (Segment Anything by Meta AI), and returned to the user on the screen. The user can hover over each region and select the region of interest (ROI) for which matching textures are to be found. (2) Categorization: The selected ROI is input to a classification model.

To suggest relevant textures belonging to the correct category of texture (Wood, Fabric, Metal, etc), a pretrained AI model is used and fine-tuned on a dataset of masks of different textures. The model is trained to classify the texture in the selected ROI into one several categories including, Wood, Fabric and Metal. (3) Texture Search: Once the texture category of the selected ROI is known, image features for the selected ROI are calculated using a pretrained image embedding model.

The calculated features are then compared against the calculated features of the textures belonging to the category predicted during the categorization step and the top k matching results are returned. The development of Toggle3D.ai's AI photo segmentation algorithm showcases the Company's commitment to delivering cutting-edge AI solutions that empowers designers, artists, marketers, and e-commerce owners with an advanced 3D texturing creative workflow, enhancing their customer experience and accelerating design and innovation.