DEFN: Dual-Encoder Fourier Group Harmonics Network for Three-Dimensional Indistinct-Boundary Object Segmentation
This addresses the problem of accurate segmentation for medical diagnostics and surgical planning, though it appears incremental as it builds on existing encoder-based architectures with specific enhancements.
The paper tackles the challenge of segmenting indistinct-boundary medical objects in 3D by introducing the DEFN network with Stochastic Defect Injection for data augmentation and Dynamic Weight Composing loss, achieving state-of-the-art performance on the OIMHS public dataset.
The precise spatial and quantitative delineation of indistinct-boundary medical objects is paramount for the accuracy of diagnostic protocols, efficacy of surgical interventions, and reliability of postoperative assessments. Despite their significance, the effective segmentation and instantaneous three-dimensional reconstruction are significantly impeded by the paucity of representative samples in available datasets and noise artifacts. To surmount these challenges, we introduced Stochastic Defect Injection (SDi) to augment the representational diversity of challenging indistinct-boundary objects within training corpora. Consequently, we propose the Dual-Encoder Fourier Group Harmonics Network (DEFN) to tailor noise filtration, amplify detailed feature recognition, and bolster representation across diverse medical imaging scenarios. By incorporating Dynamic Weight Composing (DWC) loss dynamically adjusts model's focus based on training progression, DEFN achieves SOTA performance on the OIMHS public dataset, showcasing effectiveness in indistinct boundary contexts. Source code for DEFN is available at: https://github.com/IMOP-lab/DEFN-pytorch.