Mask Branch Network: Weakly Supervised Branch Network with a Template Mask for Classifying Masses in 3D Automated Breast Ultrasound
Automated breast ultrasound (ABUS) is being rapidly utilized for screening and diagnosing breast cancer. Breast masses, including cancers shown in ABUS scans, often appear as irregular hypoechoic areas that are hard to distinguish from background shadings. We propose a novel <i>branch</i>...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-06-01
|
| Series: | Applied Sciences |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2076-3417/12/13/6332 |
| _version_ | 1797481086955028480 |
|---|---|
| author | Daekyung Kim Haesol Park Mijung Jang Kyong-Joon Lee |
| author_facet | Daekyung Kim Haesol Park Mijung Jang Kyong-Joon Lee |
| author_sort | Daekyung Kim |
| collection | DOAJ |
| description | Automated breast ultrasound (ABUS) is being rapidly utilized for screening and diagnosing breast cancer. Breast masses, including cancers shown in ABUS scans, often appear as irregular hypoechoic areas that are hard to distinguish from background shadings. We propose a novel <i>branch</i> network architecture incorporating segmentation information of masses in the training process. The branch network is integrated into neural network, providing the spatial attention effect. The branch network boosts the performance of existing classifiers, helping to learn meaningful features around the target breast mass. For the segmentation information, we leverage the existing radiology reports without additional labeling efforts. The reports, which is generated in medical image reading process, should include the characteristics of breast masses, such as shape and orientation, and a <i>template</i> mask can be created in a rule-based manner. Experimental results show that the proposed branch network with a template mask significantly improves the performance of existing classifiers. We also provide qualitative interpretation of the proposed method by visualizing the attention effect on target objects. |
| first_indexed | 2024-03-09T22:09:28Z |
| format | Article |
| id | doaj.art-7cbbee632eed4732b3280c166e1eb984 |
| institution | Directory Open Access Journal |
| issn | 2076-3417 |
| language | English |
| last_indexed | 2024-03-09T22:09:28Z |
| publishDate | 2022-06-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Applied Sciences |
| spelling | doaj.art-7cbbee632eed4732b3280c166e1eb9842023-11-23T19:34:50ZengMDPI AGApplied Sciences2076-34172022-06-011213633210.3390/app12136332Mask Branch Network: Weakly Supervised Branch Network with a Template Mask for Classifying Masses in 3D Automated Breast UltrasoundDaekyung Kim0Haesol Park1Mijung Jang2Kyong-Joon Lee3Monitor Corporation, Seoul 06628, KoreaKorea Institute of Science and Technology, Seoul 02792, KoreaDepartment of Radiology, Seoul National University Bundang Hospital, Seongnam 13620, KoreaMonitor Corporation, Seoul 06628, KoreaAutomated breast ultrasound (ABUS) is being rapidly utilized for screening and diagnosing breast cancer. Breast masses, including cancers shown in ABUS scans, often appear as irregular hypoechoic areas that are hard to distinguish from background shadings. We propose a novel <i>branch</i> network architecture incorporating segmentation information of masses in the training process. The branch network is integrated into neural network, providing the spatial attention effect. The branch network boosts the performance of existing classifiers, helping to learn meaningful features around the target breast mass. For the segmentation information, we leverage the existing radiology reports without additional labeling efforts. The reports, which is generated in medical image reading process, should include the characteristics of breast masses, such as shape and orientation, and a <i>template</i> mask can be created in a rule-based manner. Experimental results show that the proposed branch network with a template mask significantly improves the performance of existing classifiers. We also provide qualitative interpretation of the proposed method by visualizing the attention effect on target objects.https://www.mdpi.com/2076-3417/12/13/6332ultrasoundclassificationweakly supervised learning |
| spellingShingle | Daekyung Kim Haesol Park Mijung Jang Kyong-Joon Lee Mask Branch Network: Weakly Supervised Branch Network with a Template Mask for Classifying Masses in 3D Automated Breast Ultrasound Applied Sciences ultrasound classification weakly supervised learning |
| title | Mask Branch Network: Weakly Supervised Branch Network with a Template Mask for Classifying Masses in 3D Automated Breast Ultrasound |
| title_full | Mask Branch Network: Weakly Supervised Branch Network with a Template Mask for Classifying Masses in 3D Automated Breast Ultrasound |
| title_fullStr | Mask Branch Network: Weakly Supervised Branch Network with a Template Mask for Classifying Masses in 3D Automated Breast Ultrasound |
| title_full_unstemmed | Mask Branch Network: Weakly Supervised Branch Network with a Template Mask for Classifying Masses in 3D Automated Breast Ultrasound |
| title_short | Mask Branch Network: Weakly Supervised Branch Network with a Template Mask for Classifying Masses in 3D Automated Breast Ultrasound |
| title_sort | mask branch network weakly supervised branch network with a template mask for classifying masses in 3d automated breast ultrasound |
| topic | ultrasound classification weakly supervised learning |
| url | https://www.mdpi.com/2076-3417/12/13/6332 |
| work_keys_str_mv | AT daekyungkim maskbranchnetworkweaklysupervisedbranchnetworkwithatemplatemaskforclassifyingmassesin3dautomatedbreastultrasound AT haesolpark maskbranchnetworkweaklysupervisedbranchnetworkwithatemplatemaskforclassifyingmassesin3dautomatedbreastultrasound AT mijungjang maskbranchnetworkweaklysupervisedbranchnetworkwithatemplatemaskforclassifyingmassesin3dautomatedbreastultrasound AT kyongjoonlee maskbranchnetworkweaklysupervisedbranchnetworkwithatemplatemaskforclassifyingmassesin3dautomatedbreastultrasound |