A new, feasible, and convenient method based on semantic segmentation and deep learning for hemoglobin monitoring
ObjectiveNon-invasive methods for hemoglobin (Hb) monitoring can provide additional and relatively precise information between invasive measurements of Hb to help doctors' decision-making. We aimed to develop a new method for Hb monitoring based on mask R-CNN and MobileNetV3 with eye images as...
| Main Authors: | , , , , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2023-08-01
|
| Series: | Frontiers in Medicine |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/fmed.2023.1151996/full |
| _version_ | 1797693023072550912 |
|---|---|
| author | Xiao-yan Hu Yu-jie Li Xin Shu Ai-lin Song Hao Liang Yi-zhu Sun Xian-feng Wu Yong-shuai Li Li-fang Tan Zhi-yong Yang Chun-yong Yang Lin-quan Xu Yu-wen Chen Bin Yi |
| author_facet | Xiao-yan Hu Yu-jie Li Xin Shu Ai-lin Song Hao Liang Yi-zhu Sun Xian-feng Wu Yong-shuai Li Li-fang Tan Zhi-yong Yang Chun-yong Yang Lin-quan Xu Yu-wen Chen Bin Yi |
| author_sort | Xiao-yan Hu |
| collection | DOAJ |
| description | ObjectiveNon-invasive methods for hemoglobin (Hb) monitoring can provide additional and relatively precise information between invasive measurements of Hb to help doctors' decision-making. We aimed to develop a new method for Hb monitoring based on mask R-CNN and MobileNetV3 with eye images as input.MethodsSurgical patients from our center were enrolled. After image acquisition and pre-processing, the eye images, the manually selected palpebral conjunctiva, and features extracted, respectively, from the two kinds of images were used as inputs. A combination of feature engineering and regression, solely MobileNetV3, and a combination of mask R-CNN and MobileNetV3 were applied for model development. The model's performance was evaluated using metrics such as R2, explained variance score (EVS), and mean absolute error (MAE).ResultsA total of 1,065 original images were analyzed. The model's performance based on the combination of mask R-CNN and MobileNetV3 using the eye images achieved an R2, EVS, and MAE of 0.503 (95% CI, 0.499–0.507), 0.518 (95% CI, 0.515–0.522) and 1.6 g/dL (95% CI, 1.6–1.6 g/dL), which was similar to that based on MobileNetV3 using the manually selected palpebral conjunctiva images (R2: 0.509, EVS:0.516, MAE:1.6 g/dL).ConclusionWe developed a new and automatic method for Hb monitoring to help medical staffs' decision-making with high efficiency, especially in cases of disaster rescue, casualty transport, and so on. |
| first_indexed | 2024-03-12T02:37:31Z |
| format | Article |
| id | doaj.art-a463197884874a988fbf63f6fffdaa5a |
| institution | Directory Open Access Journal |
| issn | 2296-858X |
| language | English |
| last_indexed | 2024-03-12T02:37:31Z |
| publishDate | 2023-08-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Medicine |
| spelling | doaj.art-a463197884874a988fbf63f6fffdaa5a2023-09-04T14:16:51ZengFrontiers Media S.A.Frontiers in Medicine2296-858X2023-08-011010.3389/fmed.2023.11519961151996A new, feasible, and convenient method based on semantic segmentation and deep learning for hemoglobin monitoringXiao-yan Hu0Yu-jie Li1Xin Shu2Ai-lin Song3Hao Liang4Yi-zhu Sun5Xian-feng Wu6Yong-shuai Li7Li-fang Tan8Zhi-yong Yang9Chun-yong Yang10Lin-quan Xu11Yu-wen Chen12Bin Yi13Department of Anesthesiology, Southwest Hospital, Third Military Medical University (Army Medical University), Chongqing, ChinaDepartment of Anesthesiology, Southwest Hospital, Third Military Medical University (Army Medical University), Chongqing, ChinaDepartment of Anesthesiology, Southwest Hospital, Third Military Medical University (Army Medical University), Chongqing, ChinaDepartment of Anesthesiology, Southwest Hospital, Third Military Medical University (Army Medical University), Chongqing, ChinaDepartment of Anesthesiology, Southwest Hospital, Third Military Medical University (Army Medical University), Chongqing, ChinaDepartment of Anesthesiology, Southwest Hospital, Third Military Medical University (Army Medical University), Chongqing, ChinaDepartment of Anesthesiology, Southwest Hospital, Third Military Medical University (Army Medical University), Chongqing, ChinaDepartment of Anesthesiology, Southwest Hospital, Third Military Medical University (Army Medical University), Chongqing, ChinaDepartment of Anesthesiology, Southwest Hospital, Third Military Medical University (Army Medical University), Chongqing, ChinaDepartment of Anesthesiology, Southwest Hospital, Third Military Medical University (Army Medical University), Chongqing, ChinaDepartment of Anesthesiology, Southwest Hospital, Third Military Medical University (Army Medical University), Chongqing, ChinaChongqing Institute of Green and Intelligent Technology, Chinese Academy of Science, Chongqing, ChinaChongqing Institute of Green and Intelligent Technology, Chinese Academy of Science, Chongqing, ChinaDepartment of Anesthesiology, Southwest Hospital, Third Military Medical University (Army Medical University), Chongqing, ChinaObjectiveNon-invasive methods for hemoglobin (Hb) monitoring can provide additional and relatively precise information between invasive measurements of Hb to help doctors' decision-making. We aimed to develop a new method for Hb monitoring based on mask R-CNN and MobileNetV3 with eye images as input.MethodsSurgical patients from our center were enrolled. After image acquisition and pre-processing, the eye images, the manually selected palpebral conjunctiva, and features extracted, respectively, from the two kinds of images were used as inputs. A combination of feature engineering and regression, solely MobileNetV3, and a combination of mask R-CNN and MobileNetV3 were applied for model development. The model's performance was evaluated using metrics such as R2, explained variance score (EVS), and mean absolute error (MAE).ResultsA total of 1,065 original images were analyzed. The model's performance based on the combination of mask R-CNN and MobileNetV3 using the eye images achieved an R2, EVS, and MAE of 0.503 (95% CI, 0.499–0.507), 0.518 (95% CI, 0.515–0.522) and 1.6 g/dL (95% CI, 1.6–1.6 g/dL), which was similar to that based on MobileNetV3 using the manually selected palpebral conjunctiva images (R2: 0.509, EVS:0.516, MAE:1.6 g/dL).ConclusionWe developed a new and automatic method for Hb monitoring to help medical staffs' decision-making with high efficiency, especially in cases of disaster rescue, casualty transport, and so on.https://www.frontiersin.org/articles/10.3389/fmed.2023.1151996/fullcontinuous hemoglobin monitoringdeep learningsemantic segmentationmask R-CNNMobileNetV3 |
| spellingShingle | Xiao-yan Hu Yu-jie Li Xin Shu Ai-lin Song Hao Liang Yi-zhu Sun Xian-feng Wu Yong-shuai Li Li-fang Tan Zhi-yong Yang Chun-yong Yang Lin-quan Xu Yu-wen Chen Bin Yi A new, feasible, and convenient method based on semantic segmentation and deep learning for hemoglobin monitoring Frontiers in Medicine continuous hemoglobin monitoring deep learning semantic segmentation mask R-CNN MobileNetV3 |
| title | A new, feasible, and convenient method based on semantic segmentation and deep learning for hemoglobin monitoring |
| title_full | A new, feasible, and convenient method based on semantic segmentation and deep learning for hemoglobin monitoring |
| title_fullStr | A new, feasible, and convenient method based on semantic segmentation and deep learning for hemoglobin monitoring |
| title_full_unstemmed | A new, feasible, and convenient method based on semantic segmentation and deep learning for hemoglobin monitoring |
| title_short | A new, feasible, and convenient method based on semantic segmentation and deep learning for hemoglobin monitoring |
| title_sort | new feasible and convenient method based on semantic segmentation and deep learning for hemoglobin monitoring |
| topic | continuous hemoglobin monitoring deep learning semantic segmentation mask R-CNN MobileNetV3 |
| url | https://www.frontiersin.org/articles/10.3389/fmed.2023.1151996/full |
| work_keys_str_mv | AT xiaoyanhu anewfeasibleandconvenientmethodbasedonsemanticsegmentationanddeeplearningforhemoglobinmonitoring AT yujieli anewfeasibleandconvenientmethodbasedonsemanticsegmentationanddeeplearningforhemoglobinmonitoring AT xinshu anewfeasibleandconvenientmethodbasedonsemanticsegmentationanddeeplearningforhemoglobinmonitoring AT ailinsong anewfeasibleandconvenientmethodbasedonsemanticsegmentationanddeeplearningforhemoglobinmonitoring AT haoliang anewfeasibleandconvenientmethodbasedonsemanticsegmentationanddeeplearningforhemoglobinmonitoring AT yizhusun anewfeasibleandconvenientmethodbasedonsemanticsegmentationanddeeplearningforhemoglobinmonitoring AT xianfengwu anewfeasibleandconvenientmethodbasedonsemanticsegmentationanddeeplearningforhemoglobinmonitoring AT yongshuaili anewfeasibleandconvenientmethodbasedonsemanticsegmentationanddeeplearningforhemoglobinmonitoring AT lifangtan anewfeasibleandconvenientmethodbasedonsemanticsegmentationanddeeplearningforhemoglobinmonitoring AT zhiyongyang anewfeasibleandconvenientmethodbasedonsemanticsegmentationanddeeplearningforhemoglobinmonitoring AT chunyongyang anewfeasibleandconvenientmethodbasedonsemanticsegmentationanddeeplearningforhemoglobinmonitoring AT linquanxu anewfeasibleandconvenientmethodbasedonsemanticsegmentationanddeeplearningforhemoglobinmonitoring AT yuwenchen anewfeasibleandconvenientmethodbasedonsemanticsegmentationanddeeplearningforhemoglobinmonitoring AT binyi anewfeasibleandconvenientmethodbasedonsemanticsegmentationanddeeplearningforhemoglobinmonitoring AT xiaoyanhu newfeasibleandconvenientmethodbasedonsemanticsegmentationanddeeplearningforhemoglobinmonitoring AT yujieli newfeasibleandconvenientmethodbasedonsemanticsegmentationanddeeplearningforhemoglobinmonitoring AT xinshu newfeasibleandconvenientmethodbasedonsemanticsegmentationanddeeplearningforhemoglobinmonitoring AT ailinsong newfeasibleandconvenientmethodbasedonsemanticsegmentationanddeeplearningforhemoglobinmonitoring AT haoliang newfeasibleandconvenientmethodbasedonsemanticsegmentationanddeeplearningforhemoglobinmonitoring AT yizhusun newfeasibleandconvenientmethodbasedonsemanticsegmentationanddeeplearningforhemoglobinmonitoring AT xianfengwu newfeasibleandconvenientmethodbasedonsemanticsegmentationanddeeplearningforhemoglobinmonitoring AT yongshuaili newfeasibleandconvenientmethodbasedonsemanticsegmentationanddeeplearningforhemoglobinmonitoring AT lifangtan newfeasibleandconvenientmethodbasedonsemanticsegmentationanddeeplearningforhemoglobinmonitoring AT zhiyongyang newfeasibleandconvenientmethodbasedonsemanticsegmentationanddeeplearningforhemoglobinmonitoring AT chunyongyang newfeasibleandconvenientmethodbasedonsemanticsegmentationanddeeplearningforhemoglobinmonitoring AT linquanxu newfeasibleandconvenientmethodbasedonsemanticsegmentationanddeeplearningforhemoglobinmonitoring AT yuwenchen newfeasibleandconvenientmethodbasedonsemanticsegmentationanddeeplearningforhemoglobinmonitoring AT binyi newfeasibleandconvenientmethodbasedonsemanticsegmentationanddeeplearningforhemoglobinmonitoring |