The paper presents a new dataset called DAPlankton for developing and benchmarking domain adaptation methods for image recognition. The data consists of phytoplankton images captured using different imaging instruments. This dataset allows for the study of environmental aspects and provides a real-world context for developing domain adaptation methods. It comprises two subsets: DAPlankton LAB, which contains images of cultured phytoplankton, and DAPlankton SEA, which consists of images collected from the Baltic Sea. The paper also presents a benchmark comparison of three widely used domain adaptation methods.
Publication date: 8 Feb 2024
Project Page: https://arxiv.org/abs/2402.05615
Paper: https://arxiv.org/pdf/2402.05615
