Logodetect allows the detection of logos in images and videos after being trained with just one example of a logo. One-shot learning systems are not only great because they require little data, but also because they practically remove the need for specialists whenever the user wants to extend or re-purpose their functionality. This means that the benefits are manifold: the user requires little effort to collect and label training data, there is practically no time or economic costs for the training procedure, and it also provides a strategic benefit for business as they become self-sufficient for a large part of the system maintenance. Logodetect comes with pretrained models, data and an interface that allows users to detect logos in images and videos with a single command. https://github.com/Heldenkombinat/Logodetect

{"references": ["https://github.com/Heldenkombinat/Logodetect"]}