Weakly Supervised Minirhizotron Image Segmentation with MIL-CAM
This addresses a domain-specific problem for plant root imaging, offering an incremental improvement in segmentation accuracy.
The paper tackles the problem of segmenting plant roots in minirhizotron images, which are challenging due to thin roots and soil imbalance, by proposing a MIL-CAM approach that uses weak image-level labels and re-weights pixels, outperforming other methods in localization.
We present a multiple instance learning class activation map (MIL-CAM) approach for pixel-level minirhizotron image segmentation given weak image-level labels. Minirhizotrons are used to image plant roots in situ. Minirhizotron imagery is often composed of soil containing a few long and thin root objects of small diameter. The roots prove to be challenging for existing semantic image segmentation methods to discriminate. In addition to learning from weak labels, our proposed MIL-CAM approach re-weights the root versus soil pixels during analysis for improved performance due to the heavy imbalance between soil and root pixels. The proposed approach outperforms other attention map and multiple instance learning methods for localization of root objects in minirhizotron imagery.