CVision harnesses the power of AI to create video annotation solutions that enhance the speed, scale and accuracy of marine life tracking. We worked with the CVision team to design a tool that allows marine biologists and researchers to efficiently transform huge masses of video into actionable insights.
We kicked off our partnership with a day-long workshop with the CVision team. Here, we dug into the existing product, its competitors, and the needs of its audience. We interviewed a variety of marine researchers, assistants, and existing customers to further color our understanding of the product, and to uncover new ways to improve and evolve the platform.
Artificial intelligence works by learning from humans, so good tools for annotating data are critical. That’s why it was important for us to design a system that drastically streamlined workflows for customers. The more people interact with the platform, the better the data quality for an AI program to bolster and support an ever-evolving product.
One of the unique challenges for this project was to develop a process that allowed people to create their own labeling systems for annotation tasks. Our solution was a highly customizable label tree that could account for nearly any type of naming structure the user desired. Not only did this functionality make it easier for people to track their projects, it also worked to train the underlying machine-learning algorithm by first teaching it to use a specific language, and then showing it how to apply that language across the system.
As the quantity of source files grows to hundreds—sometimes thousands—of media files, so does the importance for easy access and management of them. Configurable projects, sub-sections, and dynamic filtering are ways we are currently testing within the researcher's workflow.
