Optical coherence tomography (OCT) retinal layer segmentation is a critical procedure of the modern ophthalmic process, which can be used for diagnosis and treatment of diseases such as diabetic macular edema (DME) and multiple sclerosis (MS). Due to the difficulties of low OCT image quality, highly similar retinal interlayer morphology, and the uncertain presence, shape and size of lesions, the existing algorithms do not perform well. In this work, we design an HDB-Net network for retinal layer segmentation in diseased OCT images, which solves this problem by combining global and detailed features. First, the proposed network uses a Swin transformer and Res50 as a parallel backbone network, combined with the pyramid structure in UperNet, to extract global context and aggregate multi-scale information from images. Secondly, a feature aggregation module (FAM) is designed to extract global context information from the Swin transformer and local feature information from ResNet by introducing mixed attention mechanism. Finally, the boundary awareness and feature enhancement module (BA-FEM) is used to extract the retinal layer boundary information and topological order from the low-resolution features of the shallow layer. Our approach has been validated on two public datasets, and Dice scores were 87.61% and 92.44, respectively, both outperforming other state-of-the-art technologies.
{"title":"HDB-Net: hierarchical dual-branch network for retinal layer segmentation in diseased OCT images.","authors":"Yu Chen,XueHe Zhang,Jiahui Yang,Gang Han,He Zhang,MingZhu Lai,Jie Zhao","doi":"10.1364/boe.530469","DOIUrl":"https://doi.org/10.1364/boe.530469","url":null,"abstract":"Optical coherence tomography (OCT) retinal layer segmentation is a critical procedure of the modern ophthalmic process, which can be used for diagnosis and treatment of diseases such as diabetic macular edema (DME) and multiple sclerosis (MS). Due to the difficulties of low OCT image quality, highly similar retinal interlayer morphology, and the uncertain presence, shape and size of lesions, the existing algorithms do not perform well. In this work, we design an HDB-Net network for retinal layer segmentation in diseased OCT images, which solves this problem by combining global and detailed features. First, the proposed network uses a Swin transformer and Res50 as a parallel backbone network, combined with the pyramid structure in UperNet, to extract global context and aggregate multi-scale information from images. Secondly, a feature aggregation module (FAM) is designed to extract global context information from the Swin transformer and local feature information from ResNet by introducing mixed attention mechanism. Finally, the boundary awareness and feature enhancement module (BA-FEM) is used to extract the retinal layer boundary information and topological order from the low-resolution features of the shallow layer. Our approach has been validated on two public datasets, and Dice scores were 87.61% and 92.44, respectively, both outperforming other state-of-the-art technologies.","PeriodicalId":8969,"journal":{"name":"Biomedical optics express","volume":"21 1","pages":"5359-5383"},"PeriodicalIF":3.4,"publicationDate":"2024-08-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142247638","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Accurate optical characterization of microscopic objects is crucial in academic research, product development, and clinical diagnosis. We present a method for obtaining full and high-dynamic range, angle-resolved light scattering attributes of microparticles, enabling distinction of variations in both overall morphology and detailed internal structures. This method overcomes previous limitations in observable scattering angles and dynamic range of signals through computationally assisted three-dimensional holotomography. This advancement is significant for particles spanning tens of wavelengths, such as human erythrocytes, which have historically posed measurement challenges due to faint side-scattering signals indicative of their complex interiors. Our technique addresses three key challenges in optical side-scattering analysis: limited observational angular range, reliance on simplified computational models, and low signal-to-noise ratios in both experimental and computational evaluations. We incorporate three-dimensional tomographic complex refractive index data from Fourier-transform light scattering into a tailored finite-difference time-domain simulation space. This approach facilitates precise near-to-far-field transformations. The process yields complete full-angle scattering phase functions, crucial for particles like Plasmodium falciparum-parasitized erythrocytes, predominantly involved in forward scattering. The resultant scattering data exhibit an extreme dynamic range exceeding 100 dB at various incident angles of a He-Ne laser. These findings have the potential to develop point-of-care, cost-effective, and rapid malaria diagnostic tools, inspiring further clinical and research applications in microparticle scattering.
对微观物体进行精确的光学表征对于学术研究、产品开发和临床诊断至关重要。我们提出了一种获取微颗粒的完整、高动态范围、角度分辨光散射属性的方法，从而能够区分整体形态和详细内部结构的变化。这种方法通过计算辅助三维全图成像克服了以往在可观测散射角和信号动态范围方面的限制。对于人类红细胞等波长跨度达数十个波长的颗粒来说，这一进步意义重大，因为这些颗粒内部结构复杂，会产生微弱的侧向散射信号，这在历史上曾给测量工作带来挑战。我们的技术解决了光学侧向散射分析中的三大难题：观测角度范围有限、依赖简化的计算模型以及实验和计算评估中的低信噪比。我们将傅立叶变换光散射的三维层析复折射率数据纳入定制的有限差分时域模拟空间。这种方法有助于进行精确的近场到远场转换。这一过程产生了完整的全角散射相位函数，这对恶性疟原虫寄生红细胞等主要参与前向散射的颗粒至关重要。在氦氖激光器的不同入射角度下，所得到的散射数据显示出超过 100 dB 的极限动态范围。这些发现有望开发出成本效益高、快速的疟疾床旁诊断工具，激发微粒子散射的进一步临床和研究应用。
{"title":"In silico full-angle high-dynamic range scattering of microscopic objects exploiting holotomography.","authors":"Seung Kyu Kang,Kyoohyun Kim,Jinsoo Jeong,Sunghee Hong,YongKeun Park,Jonghwa Shin","doi":"10.1364/boe.528698","DOIUrl":"https://doi.org/10.1364/boe.528698","url":null,"abstract":"Accurate optical characterization of microscopic objects is crucial in academic research, product development, and clinical diagnosis. We present a method for obtaining full and high-dynamic range, angle-resolved light scattering attributes of microparticles, enabling distinction of variations in both overall morphology and detailed internal structures. This method overcomes previous limitations in observable scattering angles and dynamic range of signals through computationally assisted three-dimensional holotomography. This advancement is significant for particles spanning tens of wavelengths, such as human erythrocytes, which have historically posed measurement challenges due to faint side-scattering signals indicative of their complex interiors. Our technique addresses three key challenges in optical side-scattering analysis: limited observational angular range, reliance on simplified computational models, and low signal-to-noise ratios in both experimental and computational evaluations. We incorporate three-dimensional tomographic complex refractive index data from Fourier-transform light scattering into a tailored finite-difference time-domain simulation space. This approach facilitates precise near-to-far-field transformations. The process yields complete full-angle scattering phase functions, crucial for particles like Plasmodium falciparum-parasitized erythrocytes, predominantly involved in forward scattering. The resultant scattering data exhibit an extreme dynamic range exceeding 100 dB at various incident angles of a He-Ne laser. These findings have the potential to develop point-of-care, cost-effective, and rapid malaria diagnostic tools, inspiring further clinical and research applications in microparticle scattering.","PeriodicalId":8969,"journal":{"name":"Biomedical optics express","volume":"14 1","pages":"5238-5250"},"PeriodicalIF":3.4,"publicationDate":"2024-08-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142268106","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Xuefei Ding,A I Hernandez-Serrano,Jacob J Young,Emma Pickwell-MacPherson
The skin, being the body's largest organ, plays a pivotal role in protecting the body against dangerous external factors. The maintenance of adequate hydration levels is essential for the skin to fulfill this protective function. However, skin hydration depends upon different biophysical factors and lifestyles, such as ethnicity, sex, age, water consumption, and many more. Consequently, methods to assess skin hydration in a precise and non-invasive manner are in high demand. In this paper, using a portable and handheld terahertz (THz) probe, we systematically examine the correlation between diverse biophysical factors and skin hydration profile in a population exceeding 300 participants. Through comparative analysis of THz light reflected from the skin against a dielectric model, we successfully extracted the thickness and hydration percentage of the outermost layer of the epidermis, the stratum corneum (SC). Our findings indicate that SC hydration and thickness are associated with variables such as daily water consumption, age, drinking coffee, and exercise. Additionally, our measurements reveal distinctions in the skin's hydration properties concerning susceptibility to UV-induced effects by bringing in the Fitzpatrick skin types. This THz-based technique holds the potential for facile integration into clinical settings for the evaluation and diagnosis of various skin-related conditions.
{"title":"Variation of skin hydration profile with biophysical factors and lifestyle revealed by in vivo terahertz sensing.","authors":"Xuefei Ding,A I Hernandez-Serrano,Jacob J Young,Emma Pickwell-MacPherson","doi":"10.1364/boe.527731","DOIUrl":"https://doi.org/10.1364/boe.527731","url":null,"abstract":"The skin, being the body's largest organ, plays a pivotal role in protecting the body against dangerous external factors. The maintenance of adequate hydration levels is essential for the skin to fulfill this protective function. However, skin hydration depends upon different biophysical factors and lifestyles, such as ethnicity, sex, age, water consumption, and many more. Consequently, methods to assess skin hydration in a precise and non-invasive manner are in high demand. In this paper, using a portable and handheld terahertz (THz) probe, we systematically examine the correlation between diverse biophysical factors and skin hydration profile in a population exceeding 300 participants. Through comparative analysis of THz light reflected from the skin against a dielectric model, we successfully extracted the thickness and hydration percentage of the outermost layer of the epidermis, the stratum corneum (SC). Our findings indicate that SC hydration and thickness are associated with variables such as daily water consumption, age, drinking coffee, and exercise. Additionally, our measurements reveal distinctions in the skin's hydration properties concerning susceptibility to UV-induced effects by bringing in the Fitzpatrick skin types. This THz-based technique holds the potential for facile integration into clinical settings for the evaluation and diagnosis of various skin-related conditions.","PeriodicalId":8969,"journal":{"name":"Biomedical optics express","volume":"10 1","pages":"5180-5198"},"PeriodicalIF":3.4,"publicationDate":"2024-08-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142247639","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The early stage of dental caries, i.e. demineralization, has always been a topic of concern to dentists. Understanding the essential mechanism of its occurrence is of great significance for the prevention and treatment of dental caries. However, owing to limitations in resolution and the detection capabilities of diagnostic tools, the study of enamel demineralization has always been a challenge. Terahertz (THz) technology, especially the combination of scanning near-field optical microscopy (s-SNOM) and THz time-domain spectroscopy (TDS), due to its nanoscale resolution, has shown great advantages in the field of biological imaging. Here, a THz s-SNOM system is used to perform near-field imaging of enamel before and after demineralization at the nanoscale. It can be found that near-field signals decrease significantly after demineralization. This is due to the changes of the crystal lattice and the transfer of mineral ions during demineralization, which leads to a decrease in the permittivity of the enamel. The novel approach in this study reveals the essence of demineralization and lays the groundwork for additional research and potential interventions.
{"title":"Unveiling enamel demineralization mechanisms by sensitive dielectric differentiation based on terahertz nanospectroscopy.","authors":"Feng Xiao,Xiaoqiuyan Zhang,Xingxing Xu,Tianyu Zhang,Fu Tang,Haowei Yin,Tao Hu,Lei Lei,Li Cheng,Min Hu","doi":"10.1364/boe.527554","DOIUrl":"https://doi.org/10.1364/boe.527554","url":null,"abstract":"The early stage of dental caries, i.e. demineralization, has always been a topic of concern to dentists. Understanding the essential mechanism of its occurrence is of great significance for the prevention and treatment of dental caries. However, owing to limitations in resolution and the detection capabilities of diagnostic tools, the study of enamel demineralization has always been a challenge. Terahertz (THz) technology, especially the combination of scanning near-field optical microscopy (s-SNOM) and THz time-domain spectroscopy (TDS), due to its nanoscale resolution, has shown great advantages in the field of biological imaging. Here, a THz s-SNOM system is used to perform near-field imaging of enamel before and after demineralization at the nanoscale. It can be found that near-field signals decrease significantly after demineralization. This is due to the changes of the crystal lattice and the transfer of mineral ions during demineralization, which leads to a decrease in the permittivity of the enamel. The novel approach in this study reveals the essence of demineralization and lays the groundwork for additional research and potential interventions.","PeriodicalId":8969,"journal":{"name":"Biomedical optics express","volume":"65 1","pages":"5229-5237"},"PeriodicalIF":3.4,"publicationDate":"2024-08-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142247640","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-08-13eCollection Date: 2024-09-01DOI: 10.1364/BOE.538890
Rachel Eimen, Mayaank Pillai, Kristen R Scarpato, Audrey K Bowden
[This corrects the article on p. 3394 in vol. 15, PMID: 38855702.].
[此处更正了第 15 卷第 3394 页的文章，PMID：38855702]。
{"title":"Erratum: Towards improved 3D reconstruction of cystoscopies through real-time feedback for frame reacquisition: erratum.","authors":"Rachel Eimen, Mayaank Pillai, Kristen R Scarpato, Audrey K Bowden","doi":"10.1364/BOE.538890","DOIUrl":"https://doi.org/10.1364/BOE.538890","url":null,"abstract":"<p><p>[This corrects the article on p. 3394 in vol. 15, PMID: 38855702.].</p>","PeriodicalId":8969,"journal":{"name":"Biomedical optics express","volume":"15 9","pages":"5227-5228"},"PeriodicalIF":2.9,"publicationDate":"2024-08-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11407264/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142280111","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Luni Hu,Rick I Cohen,Margarida Barroso,Nada N Boustany
Confocal frequency-domain fluorescence lifetime and Förster resonance energy transfer (FRET) microscopy of Chinese hamster ovary (CHO-K1) cells expressing the vinculin tension sensor (VinTS) is used to compare vinculin tension in three-dimensional (3D) multicellular aggregates and 2D cellular monolayers. In both 2D and 3D cultures, the FRET efficiency of VinTS is 5-6% lower than that of VinTL (p < 0.05), a tail-less control which cannot bind actin or paxillin. The difference between VinTS and VinTL FRET efficiency can be mitigated by treatment with the Rho-associated kinase inhibitor Y-27632, demonstrating that VinTS is under tension in both 2D and 3D cultures. However, there is an overall decrease in FRET efficiency of both VinTS and VinTL in the 3D multicellular aggregates compared with the 2D monolayers. Expression of VinTS in 2D and 3D cultures exhibits puncta consistent with cellular adhesions. While paxillin is present at the sites of VinTS expression in the 2D monolayers, it is generally absent from VinTS puncta in the 3D aggregates. The results suggest that VinTS experiences a modified environment in 3D aggregates compared with 2D monolayers and provide a basis for further investigation of molecular tension sensors in 3D tissue models.
{"title":"Comparison of vinculin tension in cellular monolayers and three-dimensional multicellular aggregates.","authors":"Luni Hu,Rick I Cohen,Margarida Barroso,Nada N Boustany","doi":"10.1364/boe.529156","DOIUrl":"https://doi.org/10.1364/boe.529156","url":null,"abstract":"Confocal frequency-domain fluorescence lifetime and Förster resonance energy transfer (FRET) microscopy of Chinese hamster ovary (CHO-K1) cells expressing the vinculin tension sensor (VinTS) is used to compare vinculin tension in three-dimensional (3D) multicellular aggregates and 2D cellular monolayers. In both 2D and 3D cultures, the FRET efficiency of VinTS is 5-6% lower than that of VinTL (p < 0.05), a tail-less control which cannot bind actin or paxillin. The difference between VinTS and VinTL FRET efficiency can be mitigated by treatment with the Rho-associated kinase inhibitor Y-27632, demonstrating that VinTS is under tension in both 2D and 3D cultures. However, there is an overall decrease in FRET efficiency of both VinTS and VinTL in the 3D multicellular aggregates compared with the 2D monolayers. Expression of VinTS in 2D and 3D cultures exhibits puncta consistent with cellular adhesions. While paxillin is present at the sites of VinTS expression in the 2D monolayers, it is generally absent from VinTS puncta in the 3D aggregates. The results suggest that VinTS experiences a modified environment in 3D aggregates compared with 2D monolayers and provide a basis for further investigation of molecular tension sensors in 3D tissue models.","PeriodicalId":8969,"journal":{"name":"Biomedical optics express","volume":"14 1","pages":"5199-5214"},"PeriodicalIF":3.4,"publicationDate":"2024-08-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142247641","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Shuai Sun,Yuanyuan Zhao,Yuxuan Hao,Pan Xue,Xiaoyang Guo,Wei Zhang,Chunxiang Zhang,Qiongdi Zhang,Wenhua Zhou,Hock Chun Ong,Xuefeng Yu,Zeren Li,Jia Li,Jianquan Yao
A rapid, sensitive and easy-to-implement approach is proposed for the detection of pathogenic nucleic acids based on phase-based plasmonic spectroscopy of metallic gratings. The plasmonic sensors were fabricated using interference lithography and functionalized with single-stranded DNA probes to specific target SARS-CoV-2. The biosensor achieved the detection of 40 fM viral nucleic acids within 5 min; furthermore, a detection capability of 1 aM (0.6 copies/µL) was acquired when combining with the recombinase polymerase amplification. Additionally, the multiplexed sensing system was demonstrated to simultaneously detect three genomic sequences on a single sensor chip, thereby enhancing diagnostic accuracy and enabling high-throughput detection.
本文提出了一种基于金属光栅相位质子光谱学的快速、灵敏和易于实施的病原体核酸检测方法。利用干涉光刻技术制作了质子传感器，并用单链 DNA 探针对特定的 SARS-CoV-2 目标进行了功能化。该生物传感器可在 5 分钟内检测到 40 fM 的病毒核酸；此外，结合重组酶聚合酶扩增，其检测能力可达 1 aM（0.6 拷贝/微升）。此外，多路复用传感系统还证明可在单个传感芯片上同时检测三种基因组序列，从而提高了诊断准确性并实现了高通量检测。
{"title":"Rapid, sensitive and multiplexed detection of SARS-CoV-2 viral nucleic acids enabled by phase-based surface plasmon resonance of metallic gratings.","authors":"Shuai Sun,Yuanyuan Zhao,Yuxuan Hao,Pan Xue,Xiaoyang Guo,Wei Zhang,Chunxiang Zhang,Qiongdi Zhang,Wenhua Zhou,Hock Chun Ong,Xuefeng Yu,Zeren Li,Jia Li,Jianquan Yao","doi":"10.1364/boe.535051","DOIUrl":"https://doi.org/10.1364/boe.535051","url":null,"abstract":"A rapid, sensitive and easy-to-implement approach is proposed for the detection of pathogenic nucleic acids based on phase-based plasmonic spectroscopy of metallic gratings. The plasmonic sensors were fabricated using interference lithography and functionalized with single-stranded DNA probes to specific target SARS-CoV-2. The biosensor achieved the detection of 40 fM viral nucleic acids within 5 min; furthermore, a detection capability of 1 aM (0.6 copies/µL) was acquired when combining with the recombinase polymerase amplification. Additionally, the multiplexed sensing system was demonstrated to simultaneously detect three genomic sequences on a single sensor chip, thereby enhancing diagnostic accuracy and enabling high-throughput detection.","PeriodicalId":8969,"journal":{"name":"Biomedical optics express","volume":"7 1","pages":"5215-5226"},"PeriodicalIF":3.4,"publicationDate":"2024-08-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142247643","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Xin Cao,Wenlong Tang,Huimin Gao,Yifan Wang,Yi Chen,Chengyi Gao,Fengjun Zhao,Linzhi Su
Cone beam X-ray luminescence computed tomography (CB-XLCT) is an emerging imaging technique with potential for early 3D tumor detection. However, the reconstruction challenge due to low light absorption and high scattering in tissues makes it a difficult inverse problem. In this study, the online dictionary learning (ODL) method, combined with iterative reduction FISTA (IR-FISTA), has been utilized to achieve high-quality reconstruction. Our method integrates IR-FISTA for efficient and accurate sparse coding, followed by an online stochastic approximation for dictionary updates, effectively capturing the sparse features inherent to the problem. Additionally, a re-sparse step is introduced to enhance the sparsity of the solution, making it better suited for CB-XLCT reconstruction. Numerical simulations and in vivo experiments were conducted to assess the performance of the method. The SODL-IR-FISTA achieved the smallest location error of 0.325 mm in in vivo experiments, which is 58% and 45% of the IVTCG-L 1 (0.562 mm) and OMP-L 0 (0.721 mm), respectively. Additionally, it has the highest DICE similarity coefficient, which is 0.748. The results demonstrate that our approach outperforms traditional methods in terms of localization precision, shape restoration, robustness, and practicality in live subjects.
{"title":"SODL-IR-FISTA: sparse online dictionary learning with iterative reduction FISTA for cone-beam X-ray luminescence computed tomography.","authors":"Xin Cao,Wenlong Tang,Huimin Gao,Yifan Wang,Yi Chen,Chengyi Gao,Fengjun Zhao,Linzhi Su","doi":"10.1364/boe.531828","DOIUrl":"https://doi.org/10.1364/boe.531828","url":null,"abstract":"Cone beam X-ray luminescence computed tomography (CB-XLCT) is an emerging imaging technique with potential for early 3D tumor detection. However, the reconstruction challenge due to low light absorption and high scattering in tissues makes it a difficult inverse problem. In this study, the online dictionary learning (ODL) method, combined with iterative reduction FISTA (IR-FISTA), has been utilized to achieve high-quality reconstruction. Our method integrates IR-FISTA for efficient and accurate sparse coding, followed by an online stochastic approximation for dictionary updates, effectively capturing the sparse features inherent to the problem. Additionally, a re-sparse step is introduced to enhance the sparsity of the solution, making it better suited for CB-XLCT reconstruction. Numerical simulations and in vivo experiments were conducted to assess the performance of the method. The SODL-IR-FISTA achieved the smallest location error of 0.325 mm in in vivo experiments, which is 58% and 45% of the IVTCG-L 1 (0.562 mm) and OMP-L 0 (0.721 mm), respectively. Additionally, it has the highest DICE similarity coefficient, which is 0.748. The results demonstrate that our approach outperforms traditional methods in terms of localization precision, shape restoration, robustness, and practicality in live subjects.","PeriodicalId":8969,"journal":{"name":"Biomedical optics express","volume":"11 1","pages":"5162-5179"},"PeriodicalIF":3.4,"publicationDate":"2024-08-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142247642","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Reactive lymphocytes may indicate diseases such as viral infections. Identifying these abnormal lymphocytes is crucial for disease diagnosis. Currently, reactive lymphocytes are mainly manually identified by pathological experts with microscopes and morphological knowledge, which is time-consuming and laborious. Some studies have used convolutional neural networks (CNNs) to identify peripheral blood leukocytes, but there are limitations in the small receptive field of the model. Our model introduces a transformer based on CNN, expands the receptive field of the model, and enables it to extract global features more efficiently. We also enhance the generalization ability of the model through virtual adversarial training (VAT) without changing the parameters of the model. Finally, our model achieves an overall accuracy of 93.66% on the test set, and the accuracy of reactive lymphocytes also reaches 88.03%. This work takes another step toward the efficient identification of reactive lymphocytes.
{"title":"Adversarial training collaborating hybrid convolution-transformer network for automatic identification of reactive lymphocytes in peripheral blood.","authors":"Liye Mei,Haoran Peng,Ping Luo,Shuangtong Jin,Hui Shen,Jing He,Wei Yang,Zhiwei Ye,Haigang Sui,Mengqing Mei,Cheng Lei,Bei Xiong","doi":"10.1364/boe.525119","DOIUrl":"https://doi.org/10.1364/boe.525119","url":null,"abstract":"Reactive lymphocytes may indicate diseases such as viral infections. Identifying these abnormal lymphocytes is crucial for disease diagnosis. Currently, reactive lymphocytes are mainly manually identified by pathological experts with microscopes and morphological knowledge, which is time-consuming and laborious. Some studies have used convolutional neural networks (CNNs) to identify peripheral blood leukocytes, but there are limitations in the small receptive field of the model. Our model introduces a transformer based on CNN, expands the receptive field of the model, and enables it to extract global features more efficiently. We also enhance the generalization ability of the model through virtual adversarial training (VAT) without changing the parameters of the model. Finally, our model achieves an overall accuracy of 93.66% on the test set, and the accuracy of reactive lymphocytes also reaches 88.03%. This work takes another step toward the efficient identification of reactive lymphocytes.","PeriodicalId":8969,"journal":{"name":"Biomedical optics express","volume":"212 1","pages":"5143-5161"},"PeriodicalIF":3.4,"publicationDate":"2024-08-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142268107","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Detecting optical defocus at the retina is crucial for accurate accommodation and emmetropization. However, the optical characteristics of ocular defocus are not fully understood. To bridge this knowledge gap, we simulated polychromatic retinal image quality by considering both the monochromatic wavefront aberrations and chromatic aberrations of the eye, both in the fovea and the periphery (nasal visual field). Our study revealed two main findings: (1) chromatic and monochromatic aberrations interact to provide a signal to the retina (chromatic optical anisotropy) to discern positive from negative defocus and (2) that chromatic optical anisotropy exhibited notable differences among refractive error groups (myopes, emmetropes and hyperopes). These findings could enhance our understanding of the underlying mechanisms of defocus detection and their subsequent implications for myopia control therapies. Further research is needed to explore the retinal architecture's ability to utilize the optical signals identified in this study.
{"title":"Chromatic cues for the sign of defocus in the peripheral retina.","authors":"Len Zheleznyak,Chang Liu,Simon Winter","doi":"10.1364/boe.537268","DOIUrl":"https://doi.org/10.1364/boe.537268","url":null,"abstract":"Detecting optical defocus at the retina is crucial for accurate accommodation and emmetropization. However, the optical characteristics of ocular defocus are not fully understood. To bridge this knowledge gap, we simulated polychromatic retinal image quality by considering both the monochromatic wavefront aberrations and chromatic aberrations of the eye, both in the fovea and the periphery (nasal visual field). Our study revealed two main findings: (1) chromatic and monochromatic aberrations interact to provide a signal to the retina (chromatic optical anisotropy) to discern positive from negative defocus and (2) that chromatic optical anisotropy exhibited notable differences among refractive error groups (myopes, emmetropes and hyperopes). These findings could enhance our understanding of the underlying mechanisms of defocus detection and their subsequent implications for myopia control therapies. Further research is needed to explore the retinal architecture's ability to utilize the optical signals identified in this study.","PeriodicalId":8969,"journal":{"name":"Biomedical optics express","volume":"13 1","pages":"5098-5114"},"PeriodicalIF":3.4,"publicationDate":"2024-08-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142247787","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","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：