Ipyannotator: the infinitely hackable annotation framework
Ítalo Epifânio, Oleksandr Pysarenko, Immanuel Bayer · The Journal of Open Source Software · 2022
Annotation is a task that associates semantic tags to a digital representation (image, video, text, and others).This process is important for supervised machine learning (ML) approaches, since the model learns, and successively improves, from data examples in form of input-output pairs.The variety of digital representations makes it difficult to develop a tool that is flexible and meets all requirements for machine learning applications in different projects.Ipyannotator is an open-source tool that allows manual annotation in multiple data formats, enabling researchers/users to explore, create, and improve their datasets without leaving their own development environment, and empowering them to extend and customize the annotation process.