SIGEDA has a large potential for use by AI developers, as it solves the issue of needing large amounts of pictures and/or videos to train convolutional neural networks (CNN). Furthermore, these synthetic images/videos provide precise 3D locations of the subject in the image, not estimated as is usually done from a photo. Also very important: the time-consuming labelling is done automatically.
The variety of datasets can be balanced as needed, avoiding bias or ethics issues, and allow for creation of images difficult to obtain, e.g. evolution over a large span of time (crop growth, symptoms of diseases).
SIGEDA used with images allows for precise detection of objects or people, while use with videos enables the detection of movements, as the dimension of time is added. An example of a past application is controlling patients in the execution of rehabilitation exercises.
Addressing target audiences and expressing needs
- Business partners – SMEs, Entrepreneurs, Large Corporations
- Business plan development
- Expanding to more markets /finding new customers
Collaborators that understand the true potential of it and help exploiting it; do the commercial work.
AI developers that want to use the tool – ITCL can generate training image datasets according requirements.
- Other Actors who can help us fulfil our market potential
- Research and Technology Organisations
- Private Investors
R&D, Technology and Innovation aspects
Further deployments in operational environments enabling market feedback and ideas for future features.
The tool has been used for AI recognition of people and different kinds of objects, it is applicable to virtually any subject.
SIGEDA is an excellent tool to avoid AI bias and can contribute to improve AI ethics. The dataset to train can be easily designed to include all kinds of varieties, e.g. people of different gender, race, age, size, etc.
SIGEDA can improve AI results in all application areas where image-based or video-based deep learning is used.
ITCL is currently seeking to develop a sustainable business model.
