Pub Date : 2022-12-06DOI: 10.1109/BIBM55620.2022.9995260
Dairong Peng, Sirui Sun, Xinyu Liu, Ju Zhou, Tong Chen
In this paper, we introduce a new biometric identity, facial tissue oxygen saturation (StO2). StO2 is an index of blood oxygen content in tissues and is related to blood vessel distribution pattern and metabolic rate. Experimental results show that classification accuracy can reach 83.33% in 42 participants with different stress states by using StO2 as the only input to the ResNet-50 model. We also proposed a module called StO2Net to eliminate the effects of stress on classification. The highest accuracy can reach up to 90.48% when the module is used. This pilot study shows that facial StO2 can be a promising biometric feature for identity recognition.
{"title":"Facial StO2: A New Promising Biometric Identity","authors":"Dairong Peng, Sirui Sun, Xinyu Liu, Ju Zhou, Tong Chen","doi":"10.1109/BIBM55620.2022.9995260","DOIUrl":"https://doi.org/10.1109/BIBM55620.2022.9995260","url":null,"abstract":"In this paper, we introduce a new biometric identity, facial tissue oxygen saturation (StO2). StO2 is an index of blood oxygen content in tissues and is related to blood vessel distribution pattern and metabolic rate. Experimental results show that classification accuracy can reach 83.33% in 42 participants with different stress states by using StO2 as the only input to the ResNet-50 model. We also proposed a module called StO2Net to eliminate the effects of stress on classification. The highest accuracy can reach up to 90.48% when the module is used. This pilot study shows that facial StO2 can be a promising biometric feature for identity recognition.","PeriodicalId":210337,"journal":{"name":"2022 IEEE International Conference on Bioinformatics and Biomedicine (BIBM)","volume":"19 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122265743","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-12-06DOI: 10.1109/BIBM55620.2022.9995574
Zuolin Cheng, Songtao Wei, Guoqiang Yu
Empowered by advances in single-cell RNA sequencing techniques (scRNA-seq), discovering new cell types or new subsets of a cell type has become an increasingly popular research interest. This type of study, by nature, requires assessment of similarity between cell groups. However, so far there is no quantitative metric for accurate and objective evaluation of such similarity; while current practice suffers from quite a few challenges including subjectivity. In this work, we propose a novel quantitative metric of single-cell-to-target-cell-type similarity, on the basis of scRNA-seq data and the signatures or differentially expressed gene (DEG) list of the target cell type. The proposed similarity score, TySim, evaluates the statistical significance of joint differential expression of the given DEGs in the cell to be tested. For this statistical test, the null distribution is established upon full consideration of complex factors causing heterogeneous sequencing efficiency of genes/cells. The design of TySim avoids the needs for clustering and for batch effect removal on cross-platform data, detouring the accompanying risks and burdens. Being the first quantitative metric of similarity to target cell type at a single-cell resolution, TySim has the potential to facilitate and enable a variety of biological studies. We validated the effectiveness of TySim and explored the possible directions of application through three example study cases of real datasets. Experimental results demonstrate TySim’s superior performance and great potential in making contributions to biological studies.
{"title":"A Single-Cell-Resolution Quantitative Metric of Similarity to a Target Cell Type for scRNA-seq Data","authors":"Zuolin Cheng, Songtao Wei, Guoqiang Yu","doi":"10.1109/BIBM55620.2022.9995574","DOIUrl":"https://doi.org/10.1109/BIBM55620.2022.9995574","url":null,"abstract":"Empowered by advances in single-cell RNA sequencing techniques (scRNA-seq), discovering new cell types or new subsets of a cell type has become an increasingly popular research interest. This type of study, by nature, requires assessment of similarity between cell groups. However, so far there is no quantitative metric for accurate and objective evaluation of such similarity; while current practice suffers from quite a few challenges including subjectivity. In this work, we propose a novel quantitative metric of single-cell-to-target-cell-type similarity, on the basis of scRNA-seq data and the signatures or differentially expressed gene (DEG) list of the target cell type. The proposed similarity score, TySim, evaluates the statistical significance of joint differential expression of the given DEGs in the cell to be tested. For this statistical test, the null distribution is established upon full consideration of complex factors causing heterogeneous sequencing efficiency of genes/cells. The design of TySim avoids the needs for clustering and for batch effect removal on cross-platform data, detouring the accompanying risks and burdens. Being the first quantitative metric of similarity to target cell type at a single-cell resolution, TySim has the potential to facilitate and enable a variety of biological studies. We validated the effectiveness of TySim and explored the possible directions of application through three example study cases of real datasets. Experimental results demonstrate TySim’s superior performance and great potential in making contributions to biological studies.","PeriodicalId":210337,"journal":{"name":"2022 IEEE International Conference on Bioinformatics and Biomedicine (BIBM)","volume":"138 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131993492","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-12-06DOI: 10.1109/BIBM55620.2022.9995455
Jinfeng Wang, Shuaihui Hang, Yong Liang, Jin Qin, Wenzhong Wang
Alzheimer’s disease (AD) is currently one of the mainstream senile diseases recognized in the world. It is the key problem how to automatically identify the early AD based on structed Magnetic Resonance Imaging (sMRI). In order to achieve accurate recognition of AD and obtain highly relevant brain lesions, an improved SVM with group L1/2 sparse regularization and smoothing function (SGL1/2) is proposed. It can achieve sparseness within the group, and approximate the non-smooth absolute value function to a smooth function. The improved model adopts a calibrated hinge to replace the hinge loss function in traditional SVM which is abbreviated as SSGL1/2. In the experiment, the proposed model is applied to different sMRI datasets for training and testing. Compared to other regularization of the non-group level and the group level, the classification accuracy of the proposed method reaches up to 96.03%. At the same time, the algorithm can point out the important brain areas in the MRI group, which has important reference value for the doctor’s predictive work.
{"title":"SSGL1/2: An Improved SVM with Smooth GroupL1/2 for Predicting AD","authors":"Jinfeng Wang, Shuaihui Hang, Yong Liang, Jin Qin, Wenzhong Wang","doi":"10.1109/BIBM55620.2022.9995455","DOIUrl":"https://doi.org/10.1109/BIBM55620.2022.9995455","url":null,"abstract":"Alzheimer’s disease (AD) is currently one of the mainstream senile diseases recognized in the world. It is the key problem how to automatically identify the early AD based on structed Magnetic Resonance Imaging (sMRI). In order to achieve accurate recognition of AD and obtain highly relevant brain lesions, an improved SVM with group L1/2 sparse regularization and smoothing function (SGL1/2) is proposed. It can achieve sparseness within the group, and approximate the non-smooth absolute value function to a smooth function. The improved model adopts a calibrated hinge to replace the hinge loss function in traditional SVM which is abbreviated as SSGL1/2. In the experiment, the proposed model is applied to different sMRI datasets for training and testing. Compared to other regularization of the non-group level and the group level, the classification accuracy of the proposed method reaches up to 96.03%. At the same time, the algorithm can point out the important brain areas in the MRI group, which has important reference value for the doctor’s predictive work.","PeriodicalId":210337,"journal":{"name":"2022 IEEE International Conference on Bioinformatics and Biomedicine (BIBM)","volume":"35 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132478351","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-12-06DOI: 10.1109/BIBM55620.2022.9995010
L. Zeng, Haoran Ma, L. Xiang, Shikui Tu, Ying Wang, Lie-bin Zhao, Lei Xu
Timely recognition of extubation readiness is critical, because prolonged and premature intubation will lead to sever complications and costs. Clinical assessment is time consuming and challenging and it has attracted increasing attention of machine learning in recent years. However, the data used for extubation predictions have the following flaws: 1) Manual recording errors and missing data; 2) Unreliable ventilation labels due to inadequate judgement from clinicians. Both may possibly lead to wrong ventilation labels, but existing machine learning methods for extubation prediction largely ignored this critical issue. In this paper, we proposed a self-rectifying deep learning method for extubation readiness prediction, called VentSR. It improves the prediction performance by a self-rectifying strategy, and the rectification is achieved through model training without clinical experience. To be detailed, VentSR firstly identifies possibly wrong samples by two components: Inconsistency between K-means and Labels (IKL) and Inconsistency between Model Predictions and Labels (IPL). IKL partitions a rough subset, and IPL iteratively refines this subset through training. Additionally, we designed Adjustment Operation to enhance IPL ability for refinement. Samples identified in this subset are rectified and used to train the model. The unrectified test set is directly fed into the trained model to obtain prediction results. Experiments demonstrate that VentSR outperforms other baselines. Further comparisons on high-confidence test set indicate that VentSR achieves 79.4 AUPRC, increasing by 26.0%. Feature importance analysis and case study illustration again reveals that VentSR are of potential practical usage of informing clinicians with accurate extubation readiness.
{"title":"VentSR: A Self-Rectifying Deep Learning Method for Extubation Readiness Prediction","authors":"L. Zeng, Haoran Ma, L. Xiang, Shikui Tu, Ying Wang, Lie-bin Zhao, Lei Xu","doi":"10.1109/BIBM55620.2022.9995010","DOIUrl":"https://doi.org/10.1109/BIBM55620.2022.9995010","url":null,"abstract":"Timely recognition of extubation readiness is critical, because prolonged and premature intubation will lead to sever complications and costs. Clinical assessment is time consuming and challenging and it has attracted increasing attention of machine learning in recent years. However, the data used for extubation predictions have the following flaws: 1) Manual recording errors and missing data; 2) Unreliable ventilation labels due to inadequate judgement from clinicians. Both may possibly lead to wrong ventilation labels, but existing machine learning methods for extubation prediction largely ignored this critical issue. In this paper, we proposed a self-rectifying deep learning method for extubation readiness prediction, called VentSR. It improves the prediction performance by a self-rectifying strategy, and the rectification is achieved through model training without clinical experience. To be detailed, VentSR firstly identifies possibly wrong samples by two components: Inconsistency between K-means and Labels (IKL) and Inconsistency between Model Predictions and Labels (IPL). IKL partitions a rough subset, and IPL iteratively refines this subset through training. Additionally, we designed Adjustment Operation to enhance IPL ability for refinement. Samples identified in this subset are rectified and used to train the model. The unrectified test set is directly fed into the trained model to obtain prediction results. Experiments demonstrate that VentSR outperforms other baselines. Further comparisons on high-confidence test set indicate that VentSR achieves 79.4 AUPRC, increasing by 26.0%. Feature importance analysis and case study illustration again reveals that VentSR are of potential practical usage of informing clinicians with accurate extubation readiness.","PeriodicalId":210337,"journal":{"name":"2022 IEEE International Conference on Bioinformatics and Biomedicine (BIBM)","volume":"21 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134230591","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-12-06DOI: 10.1109/BIBM55620.2022.9994955
Zewei Wu, Wei Ke, Cui Wang, W. Zhang, Z. Xiong
Recording activities of zebrafish is a fundamental task in biological research that aims to accurately track individuals and recover their real-world movement trajectories from multiple viewpoint videos. In this paper, we propose a novel online tracking solution based on a holistic perspective that leverages the correlation of appearance and location across views. It first reconstructs the 3D coordinates of targets frame by frame and then tracks them directly in 3D space instead of a 2D image plane. However, it is not trivial to implement such a solution which requires the association of targets across views and neighboring frames under occlusion and parallax distortion. To cope with that, we propose the view-invariant feature representation and the Kalman filter-based 3D state estimation, and combine the advantages of both to generate robust 3D trajectories. Extensive experiments on public datasets verify the efficiency and effectiveness of the approach.
{"title":"Online 3D Reconstruction of Zebrafish Behavioral Trajectories within A Holistic Perspective","authors":"Zewei Wu, Wei Ke, Cui Wang, W. Zhang, Z. Xiong","doi":"10.1109/BIBM55620.2022.9994955","DOIUrl":"https://doi.org/10.1109/BIBM55620.2022.9994955","url":null,"abstract":"Recording activities of zebrafish is a fundamental task in biological research that aims to accurately track individuals and recover their real-world movement trajectories from multiple viewpoint videos. In this paper, we propose a novel online tracking solution based on a holistic perspective that leverages the correlation of appearance and location across views. It first reconstructs the 3D coordinates of targets frame by frame and then tracks them directly in 3D space instead of a 2D image plane. However, it is not trivial to implement such a solution which requires the association of targets across views and neighboring frames under occlusion and parallax distortion. To cope with that, we propose the view-invariant feature representation and the Kalman filter-based 3D state estimation, and combine the advantages of both to generate robust 3D trajectories. Extensive experiments on public datasets verify the efficiency and effectiveness of the approach.","PeriodicalId":210337,"journal":{"name":"2022 IEEE International Conference on Bioinformatics and Biomedicine (BIBM)","volume":"23 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134506617","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-12-06DOI: 10.1109/BIBM55620.2022.9994859
Chaoran Zhang, Xiangfeng Yan, Yong Liu
Molecular property prediction has received great attention due to its wide application in biomedical field. Effective molecular representation learning is of substantial significance to facilitate molecular property prediction. In recent years, with the development of artificial intelligence technology, more and more computer scientists began to apply deep learning methods to molecular property prediction instead of traditional machine learning methods. However, these methods only utilize the SMILES sequences to learn sequence representation or use the molecular graphs to learn graph representation to predict molecular property, which fails to integrate the capabilities of both approaches in preserving molecular characteristics for further improvement. In this study, we propose a joint graph and sequence representation learning model for molecular property prediction, called PSGS. Specifically, PSGS utilizes a fusion layer to combine graph and sequence representation and capture the critical features of the molecular. In addition, PSGS is trained by a new self-supervised task, which maximizes the similarity between graph and sequence representations of the same molecular by using a pseudo-Siamese neural network. We conduct extensive experiments to compare our model with state-of-the-art models. Experimental results show that our model significantly outperforms the current state-of-the-art methods on four independent datasets.
{"title":"Pseudo-Siamese Neural Network Based Graph and Sequence Representation Learning for Molecular Property Prediction","authors":"Chaoran Zhang, Xiangfeng Yan, Yong Liu","doi":"10.1109/BIBM55620.2022.9994859","DOIUrl":"https://doi.org/10.1109/BIBM55620.2022.9994859","url":null,"abstract":"Molecular property prediction has received great attention due to its wide application in biomedical field. Effective molecular representation learning is of substantial significance to facilitate molecular property prediction. In recent years, with the development of artificial intelligence technology, more and more computer scientists began to apply deep learning methods to molecular property prediction instead of traditional machine learning methods. However, these methods only utilize the SMILES sequences to learn sequence representation or use the molecular graphs to learn graph representation to predict molecular property, which fails to integrate the capabilities of both approaches in preserving molecular characteristics for further improvement. In this study, we propose a joint graph and sequence representation learning model for molecular property prediction, called PSGS. Specifically, PSGS utilizes a fusion layer to combine graph and sequence representation and capture the critical features of the molecular. In addition, PSGS is trained by a new self-supervised task, which maximizes the similarity between graph and sequence representations of the same molecular by using a pseudo-Siamese neural network. We conduct extensive experiments to compare our model with state-of-the-art models. Experimental results show that our model significantly outperforms the current state-of-the-art methods on four independent datasets.","PeriodicalId":210337,"journal":{"name":"2022 IEEE International Conference on Bioinformatics and Biomedicine (BIBM)","volume":"27 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134157457","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-12-06DOI: 10.1109/BIBM55620.2022.9995567
Yadong Liu, Zhongyu Liu, Tao Jiang, Tianyi Zang, Yadong Wang
DNA methylation provides a pivotal layer of epigenetic regulation in eukaryotes that has significant involvement for numerous biological processes in health and disease. Recent long-read sequencing technology including Oxford Nanopore sequencing and PacBio HiFi sequencing greatly expands the capacity of long-range, single-molecule, and direct DNA modification detection from reads without extra laboratory techniques. A growing number of analytical pipelines including base-calling and 5mC methylation detection have been developed, but there is still a lack of comprehensive evaluations of the two sequencing technologies. Here, we assess the performance of different methylation-calling pipelines based on Nanopore and HiFi sequencing datasets to provide a systematic evaluation to guide researchers on how to select the long-read sequencing technologies in performing human epigenome-wide studies.
{"title":"Comparison of the Nanopore and PacBio sequencing technologies for DNA 5-methylcytosine detection","authors":"Yadong Liu, Zhongyu Liu, Tao Jiang, Tianyi Zang, Yadong Wang","doi":"10.1109/BIBM55620.2022.9995567","DOIUrl":"https://doi.org/10.1109/BIBM55620.2022.9995567","url":null,"abstract":"DNA methylation provides a pivotal layer of epigenetic regulation in eukaryotes that has significant involvement for numerous biological processes in health and disease. Recent long-read sequencing technology including Oxford Nanopore sequencing and PacBio HiFi sequencing greatly expands the capacity of long-range, single-molecule, and direct DNA modification detection from reads without extra laboratory techniques. A growing number of analytical pipelines including base-calling and 5mC methylation detection have been developed, but there is still a lack of comprehensive evaluations of the two sequencing technologies. Here, we assess the performance of different methylation-calling pipelines based on Nanopore and HiFi sequencing datasets to provide a systematic evaluation to guide researchers on how to select the long-read sequencing technologies in performing human epigenome-wide studies.","PeriodicalId":210337,"journal":{"name":"2022 IEEE International Conference on Bioinformatics and Biomedicine (BIBM)","volume":"51 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127554081","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-12-06DOI: 10.1109/BIBM55620.2022.9995476
Zepeng Li, Zhengyi An, Wenchuan Cheng, Jiawei Zhou, Bin Hu
Suicide causes serious harm to individuals, families and society, and becomes a social problem of widespread concern. Therefore, it is necessary to find and intervene individuals at risk of suicide as soon as possible. In recent years, social media data has successfully been leveraged for suicide risk detection. However, for fine-grained suicide risk detection, the existing models ignore the deviation between the predicted results and the real results when making wrong predictions, and do not pay attention to the semantic information contained in the labels. This paper proposes a deep learning model based on Label-Text Correlation and Deviation Punishment (LTC-DP). While learning the semantic relation adequately between the text and the corresponding label, the model can give different punishment adaptively according to the deviation degrees between the predicted results and the real result. The experimental results show that compared with the baseline model, the proposed model has better performance in fine-grained suicide risk detection. In addition, we release a fine-grained suicide risk detection data set based on Weibo, the data set is available at https://github.com/cxyazy/FGCSD-main.
{"title":"Using Label-text Correlation and Deviation Punishment for Fine-grained Suicide Risk Detection in Social Media","authors":"Zepeng Li, Zhengyi An, Wenchuan Cheng, Jiawei Zhou, Bin Hu","doi":"10.1109/BIBM55620.2022.9995476","DOIUrl":"https://doi.org/10.1109/BIBM55620.2022.9995476","url":null,"abstract":"Suicide causes serious harm to individuals, families and society, and becomes a social problem of widespread concern. Therefore, it is necessary to find and intervene individuals at risk of suicide as soon as possible. In recent years, social media data has successfully been leveraged for suicide risk detection. However, for fine-grained suicide risk detection, the existing models ignore the deviation between the predicted results and the real results when making wrong predictions, and do not pay attention to the semantic information contained in the labels. This paper proposes a deep learning model based on Label-Text Correlation and Deviation Punishment (LTC-DP). While learning the semantic relation adequately between the text and the corresponding label, the model can give different punishment adaptively according to the deviation degrees between the predicted results and the real result. The experimental results show that compared with the baseline model, the proposed model has better performance in fine-grained suicide risk detection. In addition, we release a fine-grained suicide risk detection data set based on Weibo, the data set is available at https://github.com/cxyazy/FGCSD-main.","PeriodicalId":210337,"journal":{"name":"2022 IEEE International Conference on Bioinformatics and Biomedicine (BIBM)","volume":"104 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131170111","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Surgery is the only viable treatment for cataract patients with visual acuity (VA) impairment. Clinically, to assess the necessity of cataract surgery, accurately predicting postoperative VA before surgery by analyzing multi-view optical coherence tomography (OCT) images is crucially needed. Unfortunately, due to complicated fundus conditions, determining postoperative VA remains difficult for medical experts. Deep learning methods for this problem were developed in recent years. Although effective, these methods still face several issues, such as not efficiently exploring potential relations between multi-view OCT images, neglecting the key role of clinical prior knowledge (e.g., preoperative VA value), and using only regression-based metrics which are lacking reference. In this paper, we propose a novel Cross-token Transformer Network (CTT-Net) for postoperative VA prediction by analyzing both the multi-view OCT images and preoperative VA. To effectively fuse multi-view features of OCT images, we develop cross-token attention that could restrict redundant/unnecessary attention flow. Further, we utilize the preoperative VA value to provide more information for postoperative VA prediction and facilitate fusion between views. Moreover, we design an auxiliary classification loss to improve model performance and assess VA recovery more sufficiently, avoiding the limitation by only using the regression metrics. To evaluate CTT-Net, we build a multi-view OCT image dataset collected from our collaborative hospital. A set of extensive experiments validate the effectiveness of our model compared to existing methods in various metrics. Code is available at: https://github.con wjh892521292/Cataract-OCT.
{"title":"CTT-Net: A Multi-view Cross-token Transformer for Cataract Postoperative Visual Acuity Prediction","authors":"Jinhong Wang, Jingwen Wang, Tingting Chen, Wenhao Zheng, Zhe Xu, Xingdi Wu, Wendeng Xu, Haochao Ying, D. Chen, Jian Wu","doi":"10.1109/BIBM55620.2022.9995392","DOIUrl":"https://doi.org/10.1109/BIBM55620.2022.9995392","url":null,"abstract":"Surgery is the only viable treatment for cataract patients with visual acuity (VA) impairment. Clinically, to assess the necessity of cataract surgery, accurately predicting postoperative VA before surgery by analyzing multi-view optical coherence tomography (OCT) images is crucially needed. Unfortunately, due to complicated fundus conditions, determining postoperative VA remains difficult for medical experts. Deep learning methods for this problem were developed in recent years. Although effective, these methods still face several issues, such as not efficiently exploring potential relations between multi-view OCT images, neglecting the key role of clinical prior knowledge (e.g., preoperative VA value), and using only regression-based metrics which are lacking reference. In this paper, we propose a novel Cross-token Transformer Network (CTT-Net) for postoperative VA prediction by analyzing both the multi-view OCT images and preoperative VA. To effectively fuse multi-view features of OCT images, we develop cross-token attention that could restrict redundant/unnecessary attention flow. Further, we utilize the preoperative VA value to provide more information for postoperative VA prediction and facilitate fusion between views. Moreover, we design an auxiliary classification loss to improve model performance and assess VA recovery more sufficiently, avoiding the limitation by only using the regression metrics. To evaluate CTT-Net, we build a multi-view OCT image dataset collected from our collaborative hospital. A set of extensive experiments validate the effectiveness of our model compared to existing methods in various metrics. Code is available at: https://github.con wjh892521292/Cataract-OCT.","PeriodicalId":210337,"journal":{"name":"2022 IEEE International Conference on Bioinformatics and Biomedicine (BIBM)","volume":"58 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132823510","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-12-06DOI: 10.1109/BIBM55620.2022.9995504
Jing Xing, H. Mouchère
Crohn’s disease is a type of inflammatory bowel illness that is typically identified v ia computer-aided diagnosis (CAD), which employs images from wireless capsule endoscopy (WCE). While deep learning has recently made significant advancements in Crohn’s disease detection, its performance is still constrained by limited labeled data. We suggest using contrastive self-supervised learning methods to address these difficulties which was barely used in detection of Crohn’s disease. Besides, we discovered that, unlike supervised learning, it is difficult to monitor contrastive self-supervised pretraining process in real time. So we propose a method for evaluating the model during contrastive pretraining (EDCP) based on the Euclidean distance of the sample representation, so that the model can be monitored during pretraining. Our comprehensive experiment results show that with contrastive self-supervised learning, better results in Crohn’s disease detection can be obtained. EDCP has also been shown to reflect the model’s training progress. Furthermore, we discovered some intriguing issues with using contrastive self-supervised learning for small dataset tasks in our experiments that merit further investigation.
{"title":"Contrastive Self-Supervised Learning on Crohn’s Disease Detection","authors":"Jing Xing, H. Mouchère","doi":"10.1109/BIBM55620.2022.9995504","DOIUrl":"https://doi.org/10.1109/BIBM55620.2022.9995504","url":null,"abstract":"Crohn’s disease is a type of inflammatory bowel illness that is typically identified v ia computer-aided diagnosis (CAD), which employs images from wireless capsule endoscopy (WCE). While deep learning has recently made significant advancements in Crohn’s disease detection, its performance is still constrained by limited labeled data. We suggest using contrastive self-supervised learning methods to address these difficulties which was barely used in detection of Crohn’s disease. Besides, we discovered that, unlike supervised learning, it is difficult to monitor contrastive self-supervised pretraining process in real time. So we propose a method for evaluating the model during contrastive pretraining (EDCP) based on the Euclidean distance of the sample representation, so that the model can be monitored during pretraining. Our comprehensive experiment results show that with contrastive self-supervised learning, better results in Crohn’s disease detection can be obtained. EDCP has also been shown to reflect the model’s training progress. Furthermore, we discovered some intriguing issues with using contrastive self-supervised learning for small dataset tasks in our experiments that merit further investigation.","PeriodicalId":210337,"journal":{"name":"2022 IEEE International Conference on Bioinformatics and Biomedicine (BIBM)","volume":"57 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133571056","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：