A Review of the Challenges in Deep Learning for Skeletal and Smooth Muscle Ultrasound Images

Deep learning has aided in the improvement of diagnosis identification, evaluation, and the interpretation of muscle ultrasound images, which may benefit clinical personnel. Muscle ultrasound images presents challenges such as low image quality due to noise, insufficient data, and different characteristics between skeletal and smooth muscles that can affect the effectiveness of deep learning results. From 2018 to 2020, deep learning has the improved solutions used to overcome these challenges; however, deep learning solutions for ultrasound images have not been compared to the conditions and strategies used to comprehend the current state of knowledge for handling skeletal and smooth muscle ultrasound images. This study aims to look at the challenges and trends of deep learning performance, especially in regard to overcoming muscle ultrasound image problems such as low image quality, muscle movement in skeletal muscles, and muscle thickness in smooth muscles. Skeletal muscle segmentation presents difficulties due to the regular movement of muscles and resulting noise, recording data through skipped connections, and modified layers required for upsampling. In skeletal muscle classification, the problems faced are area-specific, thus making a cropping strategy useful. Furthermore, there is no need to add additional layer modifications for smooth muscle segmentation as muscle thickness is the main problem in such cases.

Download Full-text

Expression of alpha 1 integrin, a laminin-collagen receptor, during myogenesis and neurogenesis in the avian embryo

Development ◽

10.1242/dev.116.3.585 ◽

1992 ◽

Vol 116 (3) ◽

pp. 585-600 ◽

Cited By ~ 2

Author(s):

J.L. Duband ◽

A.M. Belkin ◽

J. Syfrig ◽

J.P. Thiery ◽

V.E. Koteliansky

Keyword(s):

Skeletal Muscle ◽

Endothelial Cells ◽

Smooth Muscle ◽

Smooth Muscles ◽

Collagen Iv ◽

Integrin Subunit ◽

Avian Embryo ◽

Subunit Expression ◽

Capillary Endothelial Cells ◽

Collagen Receptor

In this study, we have examined the spatiotemporal distribution of the alpha 1 integrin subunit, a putative laminin and collagen receptor, in avian embryos, using immunofluorescence microscopy and immunoblotting techniques. We used an antibody raised against a gizzard 175 × 10(3) M(r) membrane protein which was described previously and which we found to be immunologically identical to the chicken alpha 1 integrin subunit. In adult avian tissues, alpha 1 integrin exhibited a very restricted pattern of expression; it was detected only in smooth muscle and in capillary endothelial cells. In the developing embryo, alpha 1 integrin subunit expression was discovered in addition to smooth muscle and capillary endothelial cells, transiently, in both central and peripheral nervous systems and in striated muscles, in association with laminin and collagen IV. alpha 1 integrin was practically absent from most epithelial tissues, including the liver, pancreas and kidney tubules, and was weakly expressed by tissues that were not associated with laminin and collagen IV. In the nervous system, alpha 1 integrin subunit expression occurred predominantly at the time of early neuronal differentiation. During skeletal muscle development, alpha 1 integrin was expressed on myogenic precursors, during myoblast migration, and in differentiating myotubes. alpha 1 integrin disappeared from skeletal muscle cells as they became contractile. In visceral and vascular smooth muscles, alpha 1 integrin appeared specifically during early smooth muscle cell differentiation and, later, was permanently expressed after cell maturation. These results indicate that (i) the expression pattern of alpha 1 integrin is consistent with a function as a laminin/collagen IV receptor; (ii) during avian development, expression of the alpha 1 integrin subunit is spatially and temporally regulated; (iii) during myogenesis and neurogenesis, expression of alpha 1 integrin is transient and correlates with cell migration and differentiation.

Download Full-text

Proteins of interstitial cells of Cajal and intestinal smooth muscle, colocalized with caveolin-1

AJP Gastrointestinal and Liver Physiology ◽

10.1152/ajpgi.00222.2004 ◽

2005 ◽

Vol 288 (3) ◽

pp. G571-G585 ◽

Cited By ~ 35

Author(s):

Woo Jung Cho ◽

E. E. Daniel

Keyword(s):

Skeletal Muscle ◽

Smooth Muscle ◽

Gap Junction ◽

Interstitial Cells Of Cajal ◽

Smooth Muscles ◽

Circular Muscle ◽

Interstitial Cells ◽

Caveolin 1 ◽

Cells Of Cajal ◽

Junction Proteins

The murine jejunum and lower esophageal sphincter (LES) were examined to determine the locations of various signaling molecules and their colocalization with caveolin-1 and one another. Caveolin-1 was present in punctate sites of the plasma membranes (PM) of all smooth muscles and diffusely in all classes of interstitial cells of Cajal (ICC; identified by c-kit immunoreactivity), ICC-myenteric plexus (MP), ICC-deep muscular plexus (DMP), ICC-serosa (ICC-S), and ICC-intramuscularis (IM). In general, all ICC also contained the L-type Ca2+ (L-Ca2+) channel, the PM Ca2+ pump, and the Na+/Ca2+ exchanger-1 localized with caveolin-1. ICC in various sites also contained Ca2+-sequestering molecules such as calreticulin and calsequestrin. Calreticulin was present also in smooth muscle, frequently in the cytosol, whereas calsequestrin was present in skeletal muscle of the esophagus. Gap junction proteins connexin-43 and -40 were present in circular muscle of jejunum but not in longitudinal muscle or in LES. In some cases, these proteins were associated with ICC-DMP. The large-conductance Ca2+-activated K+ channel was present in smooth muscle and skeletal muscle of esophagus and some ICC but was not colocalized with caveolin-1. These findings suggest that all ICC have several Ca2+-handling and -sequestering molecules, although the functions of only the L-Ca2+ channel are currently known. They also suggest that gap junction proteins are located at sites where ultrastructural gap junctions are know to exist in circular muscle of intestine but not in other smooth muscles. These findings also point to the need to evaluate the function of Ca2+ sequestration in ICC.

Download Full-text

A novel pressure-jump apparatus for the microvolume analysis of protein–ligand and protein–protein interactions: its application to nucleotide binding to skeletal-muscle and smooth-muscle myosin subfragment-1

Biochemical Journal ◽

10.1042/bj20020462 ◽

2002 ◽

Vol 366 (2) ◽

pp. 643-651 ◽

Cited By ~ 25

Author(s):

David S. PEARSON ◽

Georg HOLTERMANN ◽

Patricia ELLISON ◽

Christine CREMO ◽

Michael A. GEEVES

Keyword(s):

Skeletal Muscle ◽

Smooth Muscle ◽

Protein Interactions ◽

Muscle Protein ◽

Smooth Muscles ◽

High Volume ◽

Pressure Jump ◽

Original Design ◽

Myosin Subfragment ◽

Myosin Subfragment 1

Reactions involving proteins frequently involve large changes in volume, which allows the equilibrium position to be perturbed by changes in pressure. Rapid changes in pressure can thus be used to initiate relaxation in pressure; however, this approach is seldom used, because it requires specialized equipment. We have built a microvolume (50μl) pressure-jump apparatus, powered by a piezoelectric actuator, based on the original design of Clegg and Maxfield [(1976) Rev. Sci. Instrum. 47, 1383–1393]. This equipment can apply pressure changes of ±20MPa (maximally) in time periods as short as 80μs and follow the resulting change in fluorescence signals. The system is relatively simple to use with fast (approx. 1min) exchange of samples. In the present study, we show that this system can perturb the binding of 2′(3′)-O-(N-methylanthraniloyl)-ADP to myosin subfragment-1(S1) from skeletal and smooth muscles. The kinetic data are consistent with previous work, and in addition show that, although 2′(3′)-O-(N-methylanthraniloyl)-ADP binds with a similar affinity to both proteins, the increase in molar volume for the skeletal-muscle S1 binding to ADP is half of that for the smooth-muscle protein. This high-volume change for smooth-muscle S1 may be related to the ability of ADP to induce a 23° tilt in the tail of S1 bound to actin.

Download Full-text

Troponin C-like proteins (calmodulins) from mammalian smooth muscle and other tissues

Biochemical Journal ◽

10.1042/bj1770521 ◽

1979 ◽

Vol 177 (2) ◽

pp. 521-529 ◽

Cited By ~ 124

Author(s):

R J Grand ◽

S V Perry ◽

R A Weeks

Keyword(s):

Skeletal Muscle ◽

Smooth Muscle ◽

Troponin I ◽

Smooth Muscles ◽

Bovine Brain ◽

Troponin C ◽

Rabbit Skeletal Muscle ◽

Polyacrylamide Gels ◽

Rabbit Lung ◽

Chicken Gizzard

1. An acidic protein with properties similar to those of troponin C from rabbit skeletal muscle has been shown to be present in bovine and rabbit smooth muscles, chicken gizzard and rabbit liver, kidney and lung. 2. A simple new method involving the use of organic solvents is described for the purification of the troponin C-like proteins from various tissues. 3. The troponin C-like proteins can be distinguished from rabbit skeletal-muscle toponin C by their electrophoretic behaviour on polyacrylamide gels at pH 8.3 in the presence and absence of Ca2+. The troponin C-like proteins have been shown to form complexes with rabbit skeletal-muscle troponin I that migrate on electrophoresis in polyacrylamide gels. 4. Behaviour on electrophoresis, amino acid analysis and the patterns of CNBr digests on polyacrylamide gels indicate that the troponin C-like proteins from bovine uterus and aorta, rabbit uterus, and liver and chicken gizzard are very similar to, if not identical with, bovine brain modulator protein. 5. With bovine cardiac muscle the organic-solvent method yields a preparation consisting of roughly similar amounts of troponin C and troponin C-like protein. 6. By the isotope-dilution technique, troponin C-like protein has been shown to represent 0.42% of the total protein in rabbit uterus. 7. In homogenates of smooth muscle, rabbit lung, kidney and brain, the troponin C-like proteins form a complex with other protein (or proteins) that requires Ca2+ for its formation and that is not dissociated in 9M-urea.

Download Full-text

Classification of myositis from muscle ultrasound images using deep learning

Biomedical Signal Processing and Control ◽

10.1016/j.bspc.2021.103277 ◽

2022 ◽

Vol 71 ◽

pp. 103277

Author(s):

Emine Uçar

Keyword(s):

Deep Learning ◽

Ultrasound Images ◽

Muscle Ultrasound

Download Full-text

Calsequestrin is a component of smooth muscles: the skeletal- and cardiac-muscle isoforms are both present, although in highly variable amounts and ratios

Biochemical Journal ◽

10.1042/bj3010465 ◽

1994 ◽

Vol 301 (2) ◽

pp. 465-469 ◽

Cited By ~ 19

Author(s):

P Volpe ◽

A Martini ◽

S Furlan ◽

J Meldolesi

Keyword(s):

Skeletal Muscle ◽

Smooth Muscle ◽

Smooth Muscle Cells ◽

Vas Deferens ◽

Striated Muscle ◽

Smooth Muscles ◽

High Capacity ◽

Muscle Cells ◽

Smooth Muscle Contraction ◽

Similar Amount

Expression by smooth-muscle cells of calsequestrin (CS), the low-affinity/high-capacity Ca(2+)-binding protein of striated-muscle sarcoplasmic reticulum (SR), has been investigated in recent years with conflicting results. Here we report the purification and characterization from rat vas deferens of two CS isoforms, the first deemed skeletal muscle, the second cardiac type, on account of their N-terminal amino acids and other relevant biochemical and molecular properties. Compared with vas deferens, the smooth muscles from aorta and stomach, in that order, were found to express lower amounts of CS, whereas in the uterus and bladder the protein was not detectable. The ratio between the two CS isoforms was also variable, with the stomach and aorta predominantly expressing the skeletal-muscle type and the vas deferens expressing the two CSs in roughly similar amount. Because of the property of CSs to localize within the skeletal-muscle SR lumen not uniformly, but according to the distribution of their anchorage membrane proteins, the expression of the protein suggests the existence in smooth-muscle cells of discrete endoplasmic-reticulum areas specialized in the rapidly exchanging Ca2+ storage and release, and thus in the control of a variety of functions, including smooth-muscle contraction.

Download Full-text

Deep learning for identification of fasciculation from muscle ultrasound images

Neurology and Clinical Neuroscience ◽

10.1111/ncn3.12307 ◽

2019 ◽

Vol 7 (5) ◽

pp. 267-275

Author(s):

Hiroyuki Nodera ◽

Naoko Takamatsu ◽

Hiroki Yamazaki ◽

Ryutaro Satomi ◽

Yusuke Osaki ◽

...

Keyword(s):

Deep Learning ◽

Ultrasound Images ◽

Muscle Ultrasound

Download Full-text

Electron Probe Analysis of Muscle

Proceedings, annual meeting, Electron Microscopy Society of America ◽

10.1017/s0424820100054583 ◽

1982 ◽

Vol 40 ◽

pp. 398-401

Author(s):

A. V. Somlyo ◽

H. Shuman ◽

A. P. Somlyo

Keyword(s):

Skeletal Muscle ◽

Smooth Muscle ◽

Sarcoplasmic Reticulum ◽

Resting State ◽

Binding Sites ◽

Electron Probe ◽

Frog Skeletal Muscle ◽

Electron Probe Analysis ◽

Probe Analysis

Electron probe analysis of frozen dried cryosections of frog skeletal muscle, rabbit vascular smooth muscle and of isolated, hyperpermeab1 e rabbit cardiac myocytes has been used to determine the composition of the cytoplasm and organelles in the resting state as well as during contraction. The concentration of elements within the organelles reflects the permeabilities of the organelle membranes to the cytoplasmic ions as well as binding sites. The measurements of [Ca] in the sarcoplasmic reticulum (SR) and mitochondria at rest and during contraction, have direct bearing on their role as release and/or storage sites for Ca in situ.

Download Full-text

Deep Learning Approaches for Whiteboard Image Quality Enhancement

Color and Imaging Conference ◽

10.2352/j.imagingsci.technol.2019.63.4.040404 ◽

2019 ◽

Vol 2019 (1) ◽

pp. 360-368

Author(s):

Mekides Assefa Abebe ◽

Jon Yngve Hardeberg

Keyword(s):

Deep Learning ◽

Image Quality ◽

Image Data ◽

Quality Enhancement ◽

Network Architectures ◽

Learning Approaches ◽

Data Set ◽

Image Quality Enhancement ◽

Processing Techniques ◽

White Balancing

Different whiteboard image degradations highly reduce the legibility of pen-stroke content as well as the overall quality of the images. Consequently, different researchers addressed the problem through different image enhancement techniques. Most of the state-of-the-art approaches applied common image processing techniques such as background foreground segmentation, text extraction, contrast and color enhancements and white balancing. However, such types of conventional enhancement methods are incapable of recovering severely degraded pen-stroke contents and produce artifacts in the presence of complex pen-stroke illustrations. In order to surmount such problems, the authors have proposed a deep learning based solution. They have contributed a new whiteboard image data set and adopted two deep convolutional neural network architectures for whiteboard image quality enhancement applications. Their different evaluations of the trained models demonstrated their superior performances over the conventional methods.

Download Full-text