The Experts below are selected from a list of 321 Experts worldwide ranked by ideXlab platform

Lin Yang - One of the best experts on this subject based on the ideXlab platform.

  • Deep Learning and Convolutional Neural Networks for Medical Imaging and Clinical Informatics - Deep Learning for Muscle Pathology Image Analysis.
    Deep Learning and Convolutional Neural Networks for Medical Imaging and Clinical Informatics, 2019
    Co-Authors: Yuanpu Xie, Fuyong Xing, Fujun Liu, Lin Yang
    Abstract:

    Inflammatory myopathy (IM) is a kind of heterogeneous disease that relates to disorders of muscle functionalities. The identification of IM subtypes is critical to guide effective patient treatment since each subtype requires distinct therapy. Image analysis of hematoxylin and eosin (H&E)-stained whole-slide specimens of muscle biopsies are considered as a gold standard for effective IM diagnosis. Accurate segmentation of Perimysium plays an important role in early diagnosis of many muscle inflammation diseases. However, it remains as a challenging task due to the complex appearance of the Perimysium morphology and its ambiguity to the background area. The muscle Perimysium also exhibits strong structure spanned in the entire tissue, which makes it difficult for current local patch-based methods to capture this long-range context information. In this book chapter, we propose a novel spatial clockwork recurrent neural network (spatial CW-RNN) to address those issues. Besides Perimysium segmentation, we also introduce a fully automatic whole-slide image analysis framework for IM subtype classification using deep convolutional neural networks (DCNNs).

  • spatial clockwork recurrent neural network for muscle Perimysium segmentation
    Medical Image Computing and Computer-Assisted Intervention, 2016
    Co-Authors: Yuanpu Xie, Manish Sapkota, Zizhao Zhang, Lin Yang
    Abstract:

    Accurate segmentation of Perimysium plays an important role in early diagnosis of many muscle diseases because many diseases contain different Perimysium inflammation. However, it remains as a challenging task due to the complex appearance of the perymisum morphology and its ambiguity to the background area. The muscle Perimysium also exhibits strong structure spanned in the entire tissue, which makes it difficult for current local patch-based methods to capture this long-range context information. In this paper, we propose a novel spatial clockwork recurrent neural network (spatial CW-RNN) to address those issues. Specifically, we split the entire image into a set of non-overlapping image patches, and the semantic dependencies among them are modeled by the proposed spatial CW-RNN. Our method directly takes the 2D structure of the image into consideration and is capable of encoding the context information of the entire image into the local representation of each patch. Meanwhile, we leverage on the structured regression to assign one prediction mask rather than a single class label to each local patch, which enables both efficient training and testing. We extensively test our method for Perimysium segmentation using digitized muscle microscopy images. Experimental results demonstrate the superiority of the novel spatial CW-RNN over other existing state of the arts.

  • MICCAI (2) - Spatial Clockwork Recurrent Neural Network for Muscle Perimysium Segmentation
    Medical image computing and computer-assisted intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Inte, 2016
    Co-Authors: Yuanpu Xie, Manish Sapkota, Zizhao Zhang, Lin Yang
    Abstract:

    Accurate segmentation of Perimysium plays an important role in early diagnosis of many muscle diseases because many diseases contain different Perimysium inflammation. However, it remains as a challenging task due to the complex appearance of the perymisum morphology and its ambiguity to the background area. The muscle Perimysium also exhibits strong structure spanned in the entire tissue, which makes it difficult for current local patch-based methods to capture this long-range context information. In this paper, we propose a novel spatial clockwork recurrent neural network (spatial CW-RNN) to address those issues. Specifically, we split the entire image into a set of non-overlapping image patches, and the semantic dependencies among them are modeled by the proposed spatial CW-RNN. Our method directly takes the 2D structure of the image into consideration and is capable of encoding the context information of the entire image into the local representation of each patch. Meanwhile, we leverage on the structured regression to assign one prediction mask rather than a single class label to each local patch, which enables both efficient training and testing. We extensively test our method for Perimysium segmentation using digitized muscle microscopy images. Experimental results demonstrate the superiority of the novel spatial CW-RNN over other existing state of the arts.

  • ISBI - Automatic muscle Perimysium annotation using deep convolutional neural network
    Proceedings. IEEE International Symposium on Biomedical Imaging, 2015
    Co-Authors: Manish Sapkota, Fuyong Xing, Lin Yang
    Abstract:

    Diseased skeletal muscle expresses mononuclear cell infiltration in the regions of Perimysium. Accurate annotation or segmentation of Perimysium can help biologists and clinicians to determine individualized patient treatment and allow for reasonable prognostication. However, manual Perimysium annotation is time consuming and prone to inter-observer variations. Meanwhile, the presence of ambiguous patterns in muscle images significantly challenge many traditional automatic annotation algorithms. In this paper, we propose an automatic Perimysium annotation algorithm based on deep convolutional neural network (CNN). We formulate the automatic annotation of Perimysium in muscle images as a pixel-wise classification problem, and the CNN is trained to label each image pixel with raw RGB values of the patch centered at the pixel. The algorithm is applied to 82 diseased skeletal muscle images. We have achieved an average precision of 94% on the test dataset.

  • automatic muscle Perimysium annotation using deep convolutional neural network
    International Symposium on Biomedical Imaging, 2015
    Co-Authors: Manish Sapkota, Fuyong Xing, Hai Su, Lin Yang
    Abstract:

    Diseased skeletal muscle expresses mononuclear cell infiltration in the regions of Perimysium. Accurate annotation or segmentation of Perimysium can help biologists and clinicians to determine individualized patient treatment and allow for reasonable prognostication. However, manual Perimysium annotation is time consuming and prone to inter-observer variations. Meanwhile, the presence of ambiguous patterns in muscle images significantly challenge many traditional automatic annotation algorithms. In this paper, we propose an automatic Perimysium annotation algorithm based on deep convolutional neural network (CNN). We formulate the automatic annotation of Perimysium in muscle images as a pixel-wise classification problem, and the CNN is trained to label each image pixel with raw RGB values of the patch centered at the pixel. The algorithm is applied to 82 diseased skeletal muscle images. We have achieved an average precision of 94% on the test dataset.

Koui Takahashi - One of the best experts on this subject based on the ideXlab platform.

  • Relationships between physical and structural properties of intramuscular connective tissue and toughness of raw pork.
    Animal Science Journal, 2009
    Co-Authors: Takanori Nishimura, Suhong Fang, Jun-ichi Wakamatsu, Koui Takahashi
    Abstract:

    We studied the relationships between the shear-force value and physical and structural properties of the intramuscular connective tissue (IMCT) in six classes of porcine skeletal muscle to elucidate the contribution of IMCT to toughness of raw pork. The shear-force value of raw pork correlated significantly with that of the IMCT model prepared from each class of skeletal muscle (P < 0.05). The correlation suggested that the variable toughness of pork was caused by the mechanical strength of the endomysium and Perimysium. The thickness of the secondary Perimysium correlated significantly with the shear-force value of raw pork (P < 0.05) and with that of the IMCT model (P < 0.05). The shear-force value of raw pork correlated significantly with the total amount of collagen (P < 0.05) but not with the heat-solubility of collagen. We concluded therefore that the thickness of the secondary Perimysium determines the mechanical strength of IMCT and contributes to toughness in raw pork.

  • Structural weakening of intramuscular connective tissue during postmortem aging of pork
    Animal Science Journal, 2008
    Co-Authors: Takanori Nishimura, Suhong Fang, Jun-ichi Wakamatsu, Koui Takahashi
    Abstract:

    We studied structural changes in the endomysium and Perimysium during postmortem aging of pork using the cell-maceration/scanning electron microscope method. Immediately post mortem, endomysia sheaths that house individual muscle fibers displayed a honeycomb-like structure. The sheaths of the endomysium consisted of tightly arranged collagen fibrils in a random network. The Perimysium comprised several layers of wavy sheets made up of tightly bundled collagen fibers. While the structure of the intramuscular connective tissues remained almost unchanged up to five days post mortem, the endomysium had resolved into individual collagen fibrils, and the thick sheets of the Perimysium had separated into collagen fibers and fibrils at 8 days post mortem. These results provide direct evidence for structural weakening of the endomysium and Perimysium during postmortem aging of pork. The shear-force value of raw pork decreased rapidly within six days post mortem and then decreased slowly until 14 days post mortem. Since the rapid increase in tenderness is mainly due to structural weakening of myofibrils, we conclude that the disintegration of the endomysium and Perimysium contributes to tenderization of pork during extended postmortem aging.

  • Structural changes in endomysium and Perimysium during post-mortem aging of chicken Semitendinosus muscle—Contribution of structural weakening of intramuscular connective tissue to meat tenderization
    Meat Science, 2003
    Co-Authors: Takanori Nishimura, Koui Takahashi
    Abstract:

    Post-mortem changes in endomysium and Perimysium were investigated during aging of chicken semitendinosus muscle at 4°C. Although the shear-force value of raw meat decreased rapidly within 5 h post mortem and gradually thereafter, the solubility of collagen and the ratio of each chain of soluble collagen remained unchanged during 24 h post mortem. Light microscopic studies showed that structures of endomysium and Perimysium disintegrated into several thin sheets within 12 h post mortem, and that many gaps opened in the cross-section of endomysium and Perimysium. While endomysium and Perimysium were not stained by periodic acid Schiff reagent in fresh muscle, they were markedly stained in muscle 12 h post mortem. These results provide direct evidence for the structural weakening of endomysium and Perimysium during post-mortem aging of chicken. Therefore, we conclude that the structural weakening of the intramuscular connective tissue is closely related to tenderization of chicken.

  • Structural weakening of intramuscular connective tissue during conditioning of beef.
    Meat Science, 2002
    Co-Authors: Takanori Nishimura, Akihito Hattori, Koui Takahashi
    Abstract:

    Abstract The structural changes in intramuscular connective tissues endomysium and Perimysium during conditioning of beef were investigated using an improved technique of scanning electron microscopy. In beef conditioned for 28 days of 4°C, the endomysium resolved into individual collagen fibrils and the thick sheets of Perimysium separated into collagen fibres of 4–8 μm in diameter. These results provide direct evidence for the structural weakening of endomysium and Perimysium during conditioning. The structural changes in the intramuscular connective tissue were minimal until 10 days post mortem, but clearly observable after 14 days post mortem . Therefore, it is concluded that intramuscular connective tissue shows the effect on tenderisation of extended conditioning (2–4 weeks) of beef.

  • Structural weakening of intramuscular connective tissue during Post Mortem ageing of chicken Semitendinosus muscle
    Meat Science, 2002
    Co-Authors: Takanori Nishimura, Koui Takahashi
    Abstract:

    Abstract The structural changes in intramuscular connective tissues endomysium and Perimysium during post mortem ageing in chicken semitendinosus muscle were investigated using an improved technique of scanning electron microscopy. In post mortem chicken aged for 12 h at 4°C, the endomysium resolved into individual collagen fibrils and the perimysial sheets separated into collagen fibres. These results provide direct evidence for the structural weakening of endomysium and Perimysium during post mortem ageing. The structural changes in the intramuscular connective tissue were minimal until 6 h post mortem, but clearly observable after 12 h post mortem. It was concluded that this disintegration of the intramuscular connective tissue is the chief mechanism in tenderisation during extended post mortem ageing of chicken.

Takanori Nishimura - One of the best experts on this subject based on the ideXlab platform.

  • Relationships between physical and structural properties of intramuscular connective tissue and toughness of raw pork.
    Animal Science Journal, 2009
    Co-Authors: Takanori Nishimura, Suhong Fang, Jun-ichi Wakamatsu, Koui Takahashi
    Abstract:

    We studied the relationships between the shear-force value and physical and structural properties of the intramuscular connective tissue (IMCT) in six classes of porcine skeletal muscle to elucidate the contribution of IMCT to toughness of raw pork. The shear-force value of raw pork correlated significantly with that of the IMCT model prepared from each class of skeletal muscle (P < 0.05). The correlation suggested that the variable toughness of pork was caused by the mechanical strength of the endomysium and Perimysium. The thickness of the secondary Perimysium correlated significantly with the shear-force value of raw pork (P < 0.05) and with that of the IMCT model (P < 0.05). The shear-force value of raw pork correlated significantly with the total amount of collagen (P < 0.05) but not with the heat-solubility of collagen. We concluded therefore that the thickness of the secondary Perimysium determines the mechanical strength of IMCT and contributes to toughness in raw pork.

  • Structural weakening of intramuscular connective tissue during postmortem aging of pork
    Animal Science Journal, 2008
    Co-Authors: Takanori Nishimura, Suhong Fang, Jun-ichi Wakamatsu, Koui Takahashi
    Abstract:

    We studied structural changes in the endomysium and Perimysium during postmortem aging of pork using the cell-maceration/scanning electron microscope method. Immediately post mortem, endomysia sheaths that house individual muscle fibers displayed a honeycomb-like structure. The sheaths of the endomysium consisted of tightly arranged collagen fibrils in a random network. The Perimysium comprised several layers of wavy sheets made up of tightly bundled collagen fibers. While the structure of the intramuscular connective tissues remained almost unchanged up to five days post mortem, the endomysium had resolved into individual collagen fibrils, and the thick sheets of the Perimysium had separated into collagen fibers and fibrils at 8 days post mortem. These results provide direct evidence for structural weakening of the endomysium and Perimysium during postmortem aging of pork. The shear-force value of raw pork decreased rapidly within six days post mortem and then decreased slowly until 14 days post mortem. Since the rapid increase in tenderness is mainly due to structural weakening of myofibrils, we conclude that the disintegration of the endomysium and Perimysium contributes to tenderization of pork during extended postmortem aging.

  • Structural changes in endomysium and Perimysium during post-mortem aging of chicken Semitendinosus muscle—Contribution of structural weakening of intramuscular connective tissue to meat tenderization
    Meat Science, 2003
    Co-Authors: Takanori Nishimura, Koui Takahashi
    Abstract:

    Post-mortem changes in endomysium and Perimysium were investigated during aging of chicken semitendinosus muscle at 4°C. Although the shear-force value of raw meat decreased rapidly within 5 h post mortem and gradually thereafter, the solubility of collagen and the ratio of each chain of soluble collagen remained unchanged during 24 h post mortem. Light microscopic studies showed that structures of endomysium and Perimysium disintegrated into several thin sheets within 12 h post mortem, and that many gaps opened in the cross-section of endomysium and Perimysium. While endomysium and Perimysium were not stained by periodic acid Schiff reagent in fresh muscle, they were markedly stained in muscle 12 h post mortem. These results provide direct evidence for the structural weakening of endomysium and Perimysium during post-mortem aging of chicken. Therefore, we conclude that the structural weakening of the intramuscular connective tissue is closely related to tenderization of chicken.

  • Decorin Expression during Development of Bovine Skeletal Muscle and Its Role in Morphogenesis of the Intramuscular Connective Tissue
    Cells Tissues Organs, 2002
    Co-Authors: Takanori Nishimura, E. Futami, A. Taneichi, T. Mori, A. Hattori
    Abstract:

    Decorin is a small leucine-rich proteoglycan suspected of playing an important role in tissue morphogenesis. However, its role in the development of skeletal muscle is less clear. In the present study, the expression and spatial distribution of decorin in developing skeletal muscle of bovine fetuses were investigated, in order to provide a background for understanding the function of decorin in morphogenesis of the intramuscular connective tissue that supports muscle fibres. Western blot analysis showed that decorin already existed in skeletal muscle by 2.5 months of fetal development, and that decorin had a longer glycosaminoglycan chain in the early fetal stages than in later development, but its core protein was of the same size. Decorin mRNA was expressed at 1 month of fetal development, although its level was relatively low. Indirect immunofluorescence microscopy demonstrated that decorin was located in the Perimysium which consisted of collagen fibres, but not in the endomysium which was composed of collagen fibril networks in fetal skeletal muscle. The relatively integrated structure of the Perimysium had already formed by 2.5 months of fetal development, when muscle fibres were not tightly assembled and the surrounding endomysium was not well organized. These results suggest that decorin contributes to the formation and stabilization of collagen fibres in the Perimysium that support muscle fibres assembled with myogenesis.

  • Structural weakening of intramuscular connective tissue during conditioning of beef.
    Meat Science, 2002
    Co-Authors: Takanori Nishimura, Akihito Hattori, Koui Takahashi
    Abstract:

    Abstract The structural changes in intramuscular connective tissues endomysium and Perimysium during conditioning of beef were investigated using an improved technique of scanning electron microscopy. In beef conditioned for 28 days of 4°C, the endomysium resolved into individual collagen fibrils and the thick sheets of Perimysium separated into collagen fibres of 4–8 μm in diameter. These results provide direct evidence for the structural weakening of endomysium and Perimysium during conditioning. The structural changes in the intramuscular connective tissue were minimal until 10 days post mortem, but clearly observable after 14 days post mortem . Therefore, it is concluded that intramuscular connective tissue shows the effect on tenderisation of extended conditioning (2–4 weeks) of beef.

Yuanpu Xie - One of the best experts on this subject based on the ideXlab platform.

  • Deep Learning and Convolutional Neural Networks for Medical Imaging and Clinical Informatics - Deep Learning for Muscle Pathology Image Analysis.
    Deep Learning and Convolutional Neural Networks for Medical Imaging and Clinical Informatics, 2019
    Co-Authors: Yuanpu Xie, Fuyong Xing, Fujun Liu, Lin Yang
    Abstract:

    Inflammatory myopathy (IM) is a kind of heterogeneous disease that relates to disorders of muscle functionalities. The identification of IM subtypes is critical to guide effective patient treatment since each subtype requires distinct therapy. Image analysis of hematoxylin and eosin (H&E)-stained whole-slide specimens of muscle biopsies are considered as a gold standard for effective IM diagnosis. Accurate segmentation of Perimysium plays an important role in early diagnosis of many muscle inflammation diseases. However, it remains as a challenging task due to the complex appearance of the Perimysium morphology and its ambiguity to the background area. The muscle Perimysium also exhibits strong structure spanned in the entire tissue, which makes it difficult for current local patch-based methods to capture this long-range context information. In this book chapter, we propose a novel spatial clockwork recurrent neural network (spatial CW-RNN) to address those issues. Besides Perimysium segmentation, we also introduce a fully automatic whole-slide image analysis framework for IM subtype classification using deep convolutional neural networks (DCNNs).

  • spatial clockwork recurrent neural network for muscle Perimysium segmentation
    Medical Image Computing and Computer-Assisted Intervention, 2016
    Co-Authors: Yuanpu Xie, Manish Sapkota, Zizhao Zhang, Lin Yang
    Abstract:

    Accurate segmentation of Perimysium plays an important role in early diagnosis of many muscle diseases because many diseases contain different Perimysium inflammation. However, it remains as a challenging task due to the complex appearance of the perymisum morphology and its ambiguity to the background area. The muscle Perimysium also exhibits strong structure spanned in the entire tissue, which makes it difficult for current local patch-based methods to capture this long-range context information. In this paper, we propose a novel spatial clockwork recurrent neural network (spatial CW-RNN) to address those issues. Specifically, we split the entire image into a set of non-overlapping image patches, and the semantic dependencies among them are modeled by the proposed spatial CW-RNN. Our method directly takes the 2D structure of the image into consideration and is capable of encoding the context information of the entire image into the local representation of each patch. Meanwhile, we leverage on the structured regression to assign one prediction mask rather than a single class label to each local patch, which enables both efficient training and testing. We extensively test our method for Perimysium segmentation using digitized muscle microscopy images. Experimental results demonstrate the superiority of the novel spatial CW-RNN over other existing state of the arts.

  • MICCAI (2) - Spatial Clockwork Recurrent Neural Network for Muscle Perimysium Segmentation
    Medical image computing and computer-assisted intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Inte, 2016
    Co-Authors: Yuanpu Xie, Manish Sapkota, Zizhao Zhang, Lin Yang
    Abstract:

    Accurate segmentation of Perimysium plays an important role in early diagnosis of many muscle diseases because many diseases contain different Perimysium inflammation. However, it remains as a challenging task due to the complex appearance of the perymisum morphology and its ambiguity to the background area. The muscle Perimysium also exhibits strong structure spanned in the entire tissue, which makes it difficult for current local patch-based methods to capture this long-range context information. In this paper, we propose a novel spatial clockwork recurrent neural network (spatial CW-RNN) to address those issues. Specifically, we split the entire image into a set of non-overlapping image patches, and the semantic dependencies among them are modeled by the proposed spatial CW-RNN. Our method directly takes the 2D structure of the image into consideration and is capable of encoding the context information of the entire image into the local representation of each patch. Meanwhile, we leverage on the structured regression to assign one prediction mask rather than a single class label to each local patch, which enables both efficient training and testing. We extensively test our method for Perimysium segmentation using digitized muscle microscopy images. Experimental results demonstrate the superiority of the novel spatial CW-RNN over other existing state of the arts.

Manish Sapkota - One of the best experts on this subject based on the ideXlab platform.

  • spatial clockwork recurrent neural network for muscle Perimysium segmentation
    Medical Image Computing and Computer-Assisted Intervention, 2016
    Co-Authors: Yuanpu Xie, Manish Sapkota, Zizhao Zhang, Lin Yang
    Abstract:

    Accurate segmentation of Perimysium plays an important role in early diagnosis of many muscle diseases because many diseases contain different Perimysium inflammation. However, it remains as a challenging task due to the complex appearance of the perymisum morphology and its ambiguity to the background area. The muscle Perimysium also exhibits strong structure spanned in the entire tissue, which makes it difficult for current local patch-based methods to capture this long-range context information. In this paper, we propose a novel spatial clockwork recurrent neural network (spatial CW-RNN) to address those issues. Specifically, we split the entire image into a set of non-overlapping image patches, and the semantic dependencies among them are modeled by the proposed spatial CW-RNN. Our method directly takes the 2D structure of the image into consideration and is capable of encoding the context information of the entire image into the local representation of each patch. Meanwhile, we leverage on the structured regression to assign one prediction mask rather than a single class label to each local patch, which enables both efficient training and testing. We extensively test our method for Perimysium segmentation using digitized muscle microscopy images. Experimental results demonstrate the superiority of the novel spatial CW-RNN over other existing state of the arts.

  • MICCAI (2) - Spatial Clockwork Recurrent Neural Network for Muscle Perimysium Segmentation
    Medical image computing and computer-assisted intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Inte, 2016
    Co-Authors: Yuanpu Xie, Manish Sapkota, Zizhao Zhang, Lin Yang
    Abstract:

    Accurate segmentation of Perimysium plays an important role in early diagnosis of many muscle diseases because many diseases contain different Perimysium inflammation. However, it remains as a challenging task due to the complex appearance of the perymisum morphology and its ambiguity to the background area. The muscle Perimysium also exhibits strong structure spanned in the entire tissue, which makes it difficult for current local patch-based methods to capture this long-range context information. In this paper, we propose a novel spatial clockwork recurrent neural network (spatial CW-RNN) to address those issues. Specifically, we split the entire image into a set of non-overlapping image patches, and the semantic dependencies among them are modeled by the proposed spatial CW-RNN. Our method directly takes the 2D structure of the image into consideration and is capable of encoding the context information of the entire image into the local representation of each patch. Meanwhile, we leverage on the structured regression to assign one prediction mask rather than a single class label to each local patch, which enables both efficient training and testing. We extensively test our method for Perimysium segmentation using digitized muscle microscopy images. Experimental results demonstrate the superiority of the novel spatial CW-RNN over other existing state of the arts.

  • ISBI - Automatic muscle Perimysium annotation using deep convolutional neural network
    Proceedings. IEEE International Symposium on Biomedical Imaging, 2015
    Co-Authors: Manish Sapkota, Fuyong Xing, Lin Yang
    Abstract:

    Diseased skeletal muscle expresses mononuclear cell infiltration in the regions of Perimysium. Accurate annotation or segmentation of Perimysium can help biologists and clinicians to determine individualized patient treatment and allow for reasonable prognostication. However, manual Perimysium annotation is time consuming and prone to inter-observer variations. Meanwhile, the presence of ambiguous patterns in muscle images significantly challenge many traditional automatic annotation algorithms. In this paper, we propose an automatic Perimysium annotation algorithm based on deep convolutional neural network (CNN). We formulate the automatic annotation of Perimysium in muscle images as a pixel-wise classification problem, and the CNN is trained to label each image pixel with raw RGB values of the patch centered at the pixel. The algorithm is applied to 82 diseased skeletal muscle images. We have achieved an average precision of 94% on the test dataset.

  • automatic muscle Perimysium annotation using deep convolutional neural network
    International Symposium on Biomedical Imaging, 2015
    Co-Authors: Manish Sapkota, Fuyong Xing, Hai Su, Lin Yang
    Abstract:

    Diseased skeletal muscle expresses mononuclear cell infiltration in the regions of Perimysium. Accurate annotation or segmentation of Perimysium can help biologists and clinicians to determine individualized patient treatment and allow for reasonable prognostication. However, manual Perimysium annotation is time consuming and prone to inter-observer variations. Meanwhile, the presence of ambiguous patterns in muscle images significantly challenge many traditional automatic annotation algorithms. In this paper, we propose an automatic Perimysium annotation algorithm based on deep convolutional neural network (CNN). We formulate the automatic annotation of Perimysium in muscle images as a pixel-wise classification problem, and the CNN is trained to label each image pixel with raw RGB values of the patch centered at the pixel. The algorithm is applied to 82 diseased skeletal muscle images. We have achieved an average precision of 94% on the test dataset.