Mixed-supervised segmentation: Confidence maximization helps knowledge distillation. (arXiv:2109.10902v1 [eess.IV])

Despite achieving promising results in a breadth of medical image
segmentation tasks, deep neural networks require large training datasets with
pixel-wise annotations. Obtaining these curated datasets is a cumbersome
process which limits the application in scenarios where annotated images are
scarce. Mixed supervision is an appealing alternative for mitigating this
obstacle, where only a small fraction of the data contains complete pixel-wise
annotations and other images have a weaker form of supervision. In this work,
we propose a dual-branch architecture, where the upper branch (teacher)
receives strong annotations, while the bottom one (student) is driven by
limited supervision and guided by the upper branch. Combined with a standard
cross-entropy loss over the labeled pixels, our novel formulation integrates
two important terms: (i) a Shannon entropy loss defined over the
less-supervised images, which encourages confident student predictions in the
bottom branch; and (ii) a Kullback-Leibler (KL) divergence term, which
transfers the knowledge of the strongly supervised branch to the
less-supervised branch and guides the entropy (student-confidence) term to
avoid trivial solutions. We show that the synergy between the entropy and KL
divergence yields substantial improvements in performance. We also discuss an
interesting link between Shannon-entropy minimization and standard pseudo-mask
generation, and argue that the former should be preferred over the latter for
leveraging information from unlabeled pixels. Quantitative and qualitative
results on two publicly available datasets demonstrate that our method
significantly outperforms other strategies for semantic segmentation within a
mixed-supervision framework, as well as recent semi-supervised approaches.
Moreover, we show that the branch trained with reduced supervision and guided
by the top branch largely outperforms the latter.