A Knowledge Distillation framework for Multi-Organ Segmentation of Medaka Fish in Tomographic Image
This work addresses the challenge of organ segmentation for biological studies, offering a method to reduce annotation effort, but it is incremental as it builds on existing knowledge distillation techniques.
The paper tackles the problem of multi-organ segmentation in tomographic images of Medaka fish, which is needed for creating morphological atlases but requires large annotated datasets. The proposed self-training framework with knowledge distillation improves mean Intersection over Union by 5.9% and reduces annotation needs by three times while maintaining quality.
Morphological atlases are an important tool in organismal studies, and modern high-throughput Computed Tomography (CT) facilities can produce hundreds of full-body high-resolution volumetric images of organisms. However, creating an atlas from these volumes requires accurate organ segmentation. In the last decade, machine learning approaches have achieved incredible results in image segmentation tasks, but they require large amounts of annotated data for training. In this paper, we propose a self-training framework for multi-organ segmentation in tomographic images of Medaka fish. We utilize the pseudo-labeled data from a pretrained Teacher model and adopt a Quality Classifier to refine the pseudo-labeled data. Then, we introduce a pixel-wise knowledge distillation method to prevent overfitting to the pseudo-labeled data and improve the segmentation performance. The experimental results demonstrate that our method improves mean Intersection over Union (IoU) by 5.9% on the full dataset and enables keeping the quality while using three times less markup.