Prisma, an AI-powered photo filter company, is shifting its focus from its mobile app to SDK and API-based B2B services. Prisma announced that it will continue to support the Prisma app, but it will concentrate on offering third parties programmatic access to its various machine vision and photo editing features.
"We feel that a lot of companies need expertise in [machine vision]," Prisma CEO, Alexey Moiseenkov, told the Verge. "We can help companies put machine vision in their app because we understand how to implement the technology.
Prisma's technology empowers machines to both understand and modify image content. Prisma uses machine learning to do so, and it will now provide API and SDK access to its platform for third party access. Whether the third party wishes to implement such technology on-device, in a private cloud, or in a public instance; Prisma can provide platform access accordingly.
At launch, the features available through the API include style transfer, segmentation, object detection, scene recognition, and face landmarks. Style transfer transforms photos into art-like images (i.e. the feature that led to Prisma's mobile app popularity). Segmentation extracts foreground objects from photos and videos enabling an editor to work with the extracted images. Object detection detects individual objects within an image and categorizes accordingly. Scene recognition identifies the scene where a photo was taken. Face landmarks locate specific elements/locations on a face (e.g. contours, eyes, eyebrows, lips, etc.) which allows a user to add effects to such faces. To learn more, visit the API site.
Prisma has launched a new site, prismalabs.ai, to explain and promote its new project. The API is RESTful and SDKs are available for iOS and Android. Visit the platform site to learn more about the structure and potential of Prisma's deep learning platform.