Search papers, labs, and topics across Lattice.
This paper evaluates three data-centric methods for identifying and mitigating label noise in remote sensing datasets, a problem that significantly impacts the generalizability of deep learning models. The study systematically injects varying levels (10-70%) and types of label noise into two benchmark datasets to assess the methods' ability to filter noise and improve task performance. Results demonstrate the value of data-centric methods for both label noise identification and performance gains, providing guidance on method selection based on specific settings and objectives.
Data-centric methods can effectively identify and mitigate label noise in remote sensing data, but the best approach depends heavily on the specific noise characteristics and task objectives.
Label noise in the sense of incorrect labels is present in many real-world data sets and is known to severely limit the generalizability of deep learning models. In the field of remote sensing, however, automated treatment of label noise in data sets has received little attention to date. In particular, there is a lack of systematic analysis of the performance of data-centric methods that not only cope with label noise but also explicitly identify and isolate noisy labels. In this paper, we examine three such methods and evaluate their behavior under different label noise assumptions. To do this, we inject different types of label noise with noise levels ranging from 10 to 70% into two benchmark data sets, followed by an analysis of how well the selected methods filter the label noise and how this affects task performances. With our analyses, we clearly prove the value of data-centric methods for both parts - label noise identification and task performance improvements. Our analyses provide insights into which method is the best choice depending on the setting and objective. Finally, we show in which areas there is still a need for research in the transfer of data-centric label noise methods to remote sensing data. As such, our work is a step forward in bridging the methodological establishment of data-centric label noise methods and their usage in practical settings in the remote sensing domain.
