DEFN: Dual-Encoder Fourier Group Harmonics Network for Three-Dimensional Indistinct-Boundary Object Segmentation
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2023
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866910494833508352 |
|---|---|
| author | Jiang, Xiaohua Guo, Yihao Huang, Jian Wu, Yuting Luo, Meiyi Xu, Zhaoyang Zhang, Qianni Huang, Xingru He, Hong Jiang, Shaowei Ye, Jing Xiao, Mang |
| author_facet | Jiang, Xiaohua Guo, Yihao Huang, Jian Wu, Yuting Luo, Meiyi Xu, Zhaoyang Zhang, Qianni Huang, Xingru He, Hong Jiang, Shaowei Ye, Jing Xiao, Mang |
| contents | 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. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2311_00483 |
| institution | arXiv |
| publishDate | 2023 |
| record_format | arxiv |
| spellingShingle | DEFN: Dual-Encoder Fourier Group Harmonics Network for Three-Dimensional Indistinct-Boundary Object Segmentation Jiang, Xiaohua Guo, Yihao Huang, Jian Wu, Yuting Luo, Meiyi Xu, Zhaoyang Zhang, Qianni Huang, Xingru He, Hong Jiang, Shaowei Ye, Jing Xiao, Mang Image and Video Processing Computer Vision and Pattern Recognition 68, 92 I.4; J.3 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. |
| title | DEFN: Dual-Encoder Fourier Group Harmonics Network for Three-Dimensional Indistinct-Boundary Object Segmentation |
| topic | Image and Video Processing Computer Vision and Pattern Recognition 68, 92 I.4; J.3 |
| url | https://arxiv.org/abs/2311.00483 |