HPS-Net: Multi-Task Network for Medical Image Segmentation with Predictable Performance
In recent years, medical image segmentation (MIS) has made a huge breakthrough due to the success of deep learning. However, the existing MIS algorithms still suffer from two types of uncertainties: (1) the uncertainty of the plausible segmentation hypotheses and (2) the uncertainty of segmentation...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2021-11-01
|
| Series: | Symmetry |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2073-8994/13/11/2107 |
| _version_ | 1797508350574854144 |
|---|---|
| author | Xin Wei Huan Wan Fanghua Ye Weidong Min |
| author_facet | Xin Wei Huan Wan Fanghua Ye Weidong Min |
| author_sort | Xin Wei |
| collection | DOAJ |
| description | In recent years, medical image segmentation (MIS) has made a huge breakthrough due to the success of deep learning. However, the existing MIS algorithms still suffer from two types of uncertainties: (1) the uncertainty of the plausible segmentation hypotheses and (2) the uncertainty of segmentation performance. These two types of uncertainties affect the effectiveness of the MIS algorithm and then affect the reliability of medical diagnosis. Many studies have been done on the former but ignore the latter. Therefore, we proposed the hierarchical predictable segmentation network (HPS-Net), which consists of a new network structure, a new loss function, and a cooperative training mode. According to our knowledge, HPS-Net is the first network in the MIS area that can generate both the diverse segmentation hypotheses to avoid the uncertainty of the plausible segmentation hypotheses and the measure predictions about these hypotheses to avoid the uncertainty of segmentation performance. Extensive experiments were conducted on the LIDC-IDRI dataset and the ISIC2018 dataset. The results show that HPS-Net has the highest Dice score compared with the benchmark methods, which means it has the best segmentation performance. The results also confirmed that the proposed HPS-Net can effectively predict TNR and TPR. |
| first_indexed | 2024-03-10T05:00:51Z |
| format | Article |
| id | doaj.art-20909c15a6a84c349649bedfea3bb2aa |
| institution | Directory Open Access Journal |
| issn | 2073-8994 |
| language | English |
| last_indexed | 2024-03-10T05:00:51Z |
| publishDate | 2021-11-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Symmetry |
| spelling | doaj.art-20909c15a6a84c349649bedfea3bb2aa2023-11-23T01:45:04ZengMDPI AGSymmetry2073-89942021-11-011311210710.3390/sym13112107HPS-Net: Multi-Task Network for Medical Image Segmentation with Predictable PerformanceXin Wei0Huan Wan1Fanghua Ye2Weidong Min3School of Software, Nanchang University, 235 East Nanjing Road, Nanchang 330047, ChinaSchool of Computer Information Engineering, Jiangxi Normal University, 99 Ziyang Avenue, Nanchang 330022, ChinaSchool of Software, Nanchang University, 235 East Nanjing Road, Nanchang 330047, ChinaSchool of Software, Nanchang University, 235 East Nanjing Road, Nanchang 330047, ChinaIn recent years, medical image segmentation (MIS) has made a huge breakthrough due to the success of deep learning. However, the existing MIS algorithms still suffer from two types of uncertainties: (1) the uncertainty of the plausible segmentation hypotheses and (2) the uncertainty of segmentation performance. These two types of uncertainties affect the effectiveness of the MIS algorithm and then affect the reliability of medical diagnosis. Many studies have been done on the former but ignore the latter. Therefore, we proposed the hierarchical predictable segmentation network (HPS-Net), which consists of a new network structure, a new loss function, and a cooperative training mode. According to our knowledge, HPS-Net is the first network in the MIS area that can generate both the diverse segmentation hypotheses to avoid the uncertainty of the plausible segmentation hypotheses and the measure predictions about these hypotheses to avoid the uncertainty of segmentation performance. Extensive experiments were conducted on the LIDC-IDRI dataset and the ISIC2018 dataset. The results show that HPS-Net has the highest Dice score compared with the benchmark methods, which means it has the best segmentation performance. The results also confirmed that the proposed HPS-Net can effectively predict TNR and TPR.https://www.mdpi.com/2073-8994/13/11/2107symmetrical structuremedical imageimage segmentationdeep learningCNNsloss function |
| spellingShingle | Xin Wei Huan Wan Fanghua Ye Weidong Min HPS-Net: Multi-Task Network for Medical Image Segmentation with Predictable Performance Symmetry symmetrical structure medical image image segmentation deep learning CNNs loss function |
| title | HPS-Net: Multi-Task Network for Medical Image Segmentation with Predictable Performance |
| title_full | HPS-Net: Multi-Task Network for Medical Image Segmentation with Predictable Performance |
| title_fullStr | HPS-Net: Multi-Task Network for Medical Image Segmentation with Predictable Performance |
| title_full_unstemmed | HPS-Net: Multi-Task Network for Medical Image Segmentation with Predictable Performance |
| title_short | HPS-Net: Multi-Task Network for Medical Image Segmentation with Predictable Performance |
| title_sort | hps net multi task network for medical image segmentation with predictable performance |
| topic | symmetrical structure medical image image segmentation deep learning CNNs loss function |
| url | https://www.mdpi.com/2073-8994/13/11/2107 |
| work_keys_str_mv | AT xinwei hpsnetmultitasknetworkformedicalimagesegmentationwithpredictableperformance AT huanwan hpsnetmultitasknetworkformedicalimagesegmentationwithpredictableperformance AT fanghuaye hpsnetmultitasknetworkformedicalimagesegmentationwithpredictableperformance AT weidongmin hpsnetmultitasknetworkformedicalimagesegmentationwithpredictableperformance |