FotoInMotion
Repurposing and enriching images for immersive storytelling through smart digital tools Artificial Intelligence
Description
FotoInMotion was a project born with the purpose of transforming a single photograph into a high quality video with dynamic effects for storytelling and branding. The FotoInMotion desktop & mobile tool allows both creative content producers and the creative public to efficiently produce high-quality videos from still photography, relying on AI-driven automated editing functions and dynamic effects. Automated outputs can be easily shared on social media and/or interfaced with other software for additional editing layers.FotoInMotion’s photo post-production and automated video-editing processes will apply motion and 3D effects to high-quality still photography, inviting users to select their best images to be animated with a range of specifically designed artistic effects to engage the viewer.
Scientific Advances
FotoInMotion aimed at introducing AI approaches into the creative pipeline. It advanced the state of the art by developing machine learning models not requiring high computational resources and by fusing different types and source of metadata to enhance the personalization and context-acquisition to help the automatic creative process.
Results
Several results can be identified:
- new ML models optimized for the FotoinMotion scenarios, showing results that surpass the state of the art
- a new metadata model enabling the semantic homogenization of the media and metadata representation and packaging in the normalization of the interaction between different pipeline components.
- a ML-assisted annotation tool for the creation of datasets
- an integrated pipeline, from acquisition until publishing of media content.
Scientific Advances
FotoInMotion aimed at introducing AI approaches into the creative pipeline. It advanced the state of the art by developing machine learning models not requiring high computational resources and by fusing different types and source of metadata to enhance the personalization and context-acquisition to help the automatic creative process.
Results
Several results can be identified:
- new ML models optimized for the FotoinMotion scenarios, showing results that surpass the state of the art
- a new metadata model enabling the semantic homogenization of the media and metadata representation and packaging in the normalization of the interaction between different pipeline components.
- a ML-assisted annotation tool for the creation of datasets
- an integrated pipeline, from acquisition until publishing of media content.
