Color instance segmentation and classification of cervix images

2019 ◽  
Author(s):  
Marwa Said ◽  
Mohamed Moustafa ◽  
Ayman Wahba
Keyword(s):  
Instance Segmentation

scholarly journals HistoFlow: Label-Efficient and Interactive Deep Learning Cell Analysis

2020 ◽  
Author(s):  
Tim Henning ◽  
Benjamin Bergner ◽  
Christoph Lippert
Keyword(s):  
Machine Learning ◽  
Deep Learning ◽  
User Interface ◽  
Network Architecture ◽  
Training Data ◽  
Cell Segmentation ◽  
Annotation Tool ◽  
Cell Analysis ◽  
Instance Segmentation

Instance segmentation is a common task in quantitative cell analysis. While there are many approaches doing this using machine learning, typically, the training process requires a large amount of manually annotated data. We present HistoFlow, a software for annotation-efficient training of deep learning models for cell segmentation and analysis with an interactive user interface.It provides an assisted annotation tool to quickly draw and correct cell boundaries and use biomarkers as weak annotations. It also enables the user to create artificial training data to lower the labeling effort. We employ a universal U-Net neural network architecture that allows accurate instance segmentation and the classification of phenotypes in only a single pass of the network. Transfer learning is available through the user interface to adapt trained models to new tissue types.We demonstrate HistoFlow for fluorescence breast cancer images. The models trained using only artificial data perform comparably to those trained with time-consuming manual annotations. They outperform traditional cell segmentation algorithms and match state-of-the-art machine learning approaches. A user test shows that cells can be annotated six times faster than without the assistance of our annotation tool. Extending a segmentation model for classification of epithelial cells can be done using only 50 to 1500 annotations.Our results show that, unlike previous assumptions, it is possible to interactively train a deep learning model in a matter of minutes without many manual annotations.

scholarly journals Automatic Tooth Segmentation and Classification in Dental Panoramic X-ray Images

2020 ◽  
Author(s):  
Shuxu Zhao ◽  
QING LUO ◽  
Changrong Liu
Keyword(s):  
Recognition Task ◽  
Recall Rate ◽  
X Ray ◽  
Learning Framework ◽  
Tooth Position ◽  
Data Engineering ◽  
Segmentation Task ◽  
Instance Segmentation ◽  
Tooth Segmentation

Abstract Background: The information of tooth shape, type and tooth position plays an important role in the understanding of pathological features in dental X-ray films. It is of great significance to realize the accurate tooth segmentation and tooth classification of dental panoramic X-ray images for the construction of an intelligent dental diagnosis system.At present, the segmentation results of teeth are relatively rough, and most methods realize tooth recognition and segmentation as independent tasks, ignoring the parameter sharing between the two tasks. Therefore, an instance segmentation method which can realize tooth recognition and tooth segmentation at the same time is proposed. Methods: In model designing, the Mask R-CNN, an instance segmentation model , is adopted, which includes classification branches and segmentation branches. The classification branch can be used to complete the tooth recognition task and the segmentation branch to complete the tooth segmentation task. On this basis, the U-Net architecture is integrated to modify the segmentation branch to improve the segmentation effect. In data engineering, two classification schemes are designed, one according to the function of teeth, the other according to the position of teeth. Results: Based on the data of 400 panoramic X-ray films of teeth, we combined migration learning to conduct experiments on the TensorFlow deep learning framework. The experimental results show that compared with other methods, the classification and segmentation of teeth can be realized simultaneously in this paper, with an accuracy of more than 90%. Compared with the original model, the improved Mask R-CNN proposed in this paper improves the segmentation recall rate by 10%. In the proposed classification scheme, the accuracy of classification based on tooth function is 3% higher than that based on tooth position.Conclusions: The model proposed in this paper combines the two tasks of classification and segmentation, avoids the repetitive training of the model, and improves the segmentation precision with the improved segmentation branch. Compared with the recall rate traditional methods of tooth function classification, the proposed method based on tooth function has better classification effect.

scholarly journals COMPARATIVE ANALYSIS OF DEEP LEARNING ARCHITECTURES FOR GRAPE CLUSTER INSTANCE SEGMENTATION

2021 ◽  
Vol 9 (1) ◽  
pp. 344-352
Author(s):  
Ms. Dhanashree Barbole, Dr. Parul Jadhav
Keyword(s):  
Network Models ◽  
Testing Strategy ◽  
Neural Network Models ◽  
Cluster Identification ◽  
Proposed Model ◽  
Weight Prediction ◽  
Comparative Results ◽  
Learning Architectures ◽  
Instance Segmentation

The grape cluster identification and its segmentation for the sake of total weight prediction task of wine yard shows the need of segmentation atomization with better accuracy. The challenge of grape cluster segmentation is considered to provide solution using deep neural network models such as YOLO v3, Mask RCNN, U-net. This paper contributes in terms of the modified U-net model for the segmentation of grape clusters using training and testing strategy for the validation of the results. The results are obtained for the accuracy of the classification of pixels as part of grape cluster or outside of clusters and comparative results show improvement in segmentation using modified U-net. The accuracy, precision and recall analysis is performed and comparatively proposed model shows better results

scholarly journals A Deep Learning Instance Segmentation Approach for Global Glomerulosclerosis Assessment in Donor Kidney Biopsies

Electronics ◽  
2020 ◽  
Vol 9 (11) ◽  
pp. 1768
Author(s):  
Nicola Altini ◽  
Giacomo Donato Cascarano ◽  
Antonio Brunetti ◽  
Irio De Feudis ◽  
Domenico Buongiorno ◽  
...  
Keyword(s):  
Deep Learning ◽  
Semantic Segmentation ◽  
Point Of View ◽  
University Hospital ◽  
Kidney Section ◽  
Global Glomerulosclerosis ◽  
Bounding Boxes ◽  
F Measure ◽  
Instance Segmentation

The histological assessment of glomeruli is fundamental for determining if a kidney is suitable for transplantation. The Karpinski score is essential to evaluate the need for a single or dual kidney transplant and includes the ratio between the number of sclerotic glomeruli and the overall number of glomeruli in a kidney section. The manual evaluation of kidney biopsies performed by pathologists is time-consuming and error-prone, so an automatic framework to delineate all the glomeruli present in a kidney section can be very useful. Our experiments have been conducted on a dataset provided by the Department of Emergency and Organ Transplantations (DETO) of Bari University Hospital. This dataset is composed of 26 kidney biopsies coming from 19 donors. The rise of Convolutional Neural Networks (CNNs) has led to a realm of methods which are widely applied in Medical Imaging. Deep learning techniques are also very promising for the segmentation of glomeruli, with a variety of existing approaches. Many methods only focus on semantic segmentation—which consists in segmentation of individual pixels—or ignore the problem of discriminating between non-sclerotic and sclerotic glomeruli, so these approaches are not optimal or inadequate for transplantation assessment. In this work, we employed an end-to-end fully automatic approach based on Mask R-CNN for instance segmentation and classification of glomeruli. We also compared the results obtained with a baseline based on Faster R-CNN, which only allows detection at bounding boxes level. With respect to the existing literature, we improved the Mask R-CNN approach in sliding window contexts, by employing a variant of the Non-Maximum Suppression (NMS) algorithm, which we called Non-Maximum-Area Suppression (NMAS). The obtained results are very promising, leading to improvements over existing literature. The baseline Faster R-CNN-based approach obtained an F-Measure of 0.904 and 0.667 for non-sclerotic and sclerotic glomeruli, respectively. The Mask R-CNN approach has a significant improvement over the baseline, obtaining an F-Measure of 0.925 and 0.777 for non-sclerotic and sclerotic glomeruli, respectively. The proposed method is very promising for the instance segmentation and classification of glomeruli, and allows to make a robust evaluation of global glomerulosclerosis. We also compared Karpinski score obtained with our algorithm to that obtained with pathologists’ annotations to show the soundness of the proposed workflow from a clinical point of view.

Note on the spectral classification of carbon stars in the photographic infrared region

1966 ◽  
Vol 24 ◽  
pp. 21-23
Author(s):  
Y. Fujita
Keyword(s):  
Infrared Region ◽  
Spectral Classification ◽  
Carbon Stars

We have investigated the spectrograms (dispersion: 8Å/mm) in the photographic infrared region fromλ7500 toλ9000 of some carbon stars obtained by the coudé spectrograph of the 74-inch reflector attached to the Okayama Astrophysical Observatory. The names of the stars investigated are listed in Table 1.

Diagnostic Value of Electron Microscopy in Soft Tissue Tumors

1970 ◽  
Vol 28 ◽  
pp. 220-221
Author(s):  
Gerald Fine ◽  
Azorides R. Morales
Keyword(s):  
Cardiac Myxoma ◽  
Osteogenic Sarcoma ◽  
Diagnostic Value ◽  
Soft Tissue Tumors ◽  
Amelanotic Melanoma ◽  
Growth Patterns ◽  
Light Microscopic Level ◽  
Mesenchymal Tumors ◽  
Good Agreement

For years the separation of carcinoma and sarcoma and the subclassification of sarcomas has been based on the appearance of the tumor cells and their microscopic growth pattern and information derived from certain histochemical and special stains. Although this method of study has produced good agreement among pathologists in the separation of carcinoma from sarcoma, it has given less uniform results in the subclassification of sarcomas. There remain examples of neoplasms of different histogenesis, the classification of which is questionable because of similar cytologic and growth patterns at the light microscopic level; i.e. amelanotic melanoma versus carcinoma and occasionally sarcoma, sarcomas with an epithelial pattern of growth simulating carcinoma, histologically similar mesenchymal tumors of different histogenesis (histiocytoma versus rhabdomyosarcoma, lytic osteogenic sarcoma versus rhabdomyosarcoma), and myxomatous mesenchymal tumors of diverse histogenesis (myxoid rhabdo and liposarcomas, cardiac myxoma, myxoid neurofibroma, etc.)

Ultrastructure of mesotheliomas

1984 ◽  
Vol 42 ◽  
pp. 98-101
Author(s):  
Irving Dardick
Keyword(s):  
Electron Microscopy ◽  
Latent Period ◽  
Spindle Cell ◽  
Spindle Cell Sarcoma ◽  
Metastatic Adenocarcinoma ◽  
Cell Sarcoma ◽  
Cytological Features ◽  
Industrial Use ◽  
Degree Of Differentiation

With the extensive industrial use of asbestos in this century and the long latent period (20-50 years) between exposure and tumor presentation, the incidence of malignant mesothelioma is now increasing. Thus, surgical pathologists are more frequently faced with the dilemma of differentiating mesothelioma from metastatic adenocarcinoma and spindle-cell sarcoma involving serosal surfaces. Electron microscopy is amodality useful in clarifying this problem.In utilizing ultrastructural features in the diagnosis of mesothelioma, it is essential to appreciate that the classification of this tumor reflects a variety of morphologic forms of differing biologic behavior (Table 1). Furthermore, with the variable histology and degree of differentiation in mesotheliomas it might be expected that the ultrastructure of such tumors also reflects a range of cytological features. Such is the case.

Interpreting HVEM of muscle-cell impulse networks

1992 ◽  
Vol 50 (1) ◽  
pp. 126-127
Author(s):  
Paul DeCosta ◽  
Kyugon Cho ◽  
Stephen Shemlon ◽  
Heesung Jun ◽  
Stanley M. Dunn
Keyword(s):  
Structural Analysis ◽  
Muscle Cell ◽  
Electron Micrographs ◽  
Relative Size ◽  
Automatic Classification ◽  
Nerve Fibers ◽  
Analysis Algorithm ◽  
Structural Networks

Introduction: The analysis and interpretation of electron micrographs of cells and tissues, often requires the accurate extraction of structural networks, which either provide immediate 2D or 3D information, or from which the desired information can be inferred. The images of these structures contain lines and/or curves whose orientation, lengths, and intersections characterize the overall network.Some examples exist of studies that have been done in the analysis of networks of natural structures. In, Sebok and Roemer determine the complexity of nerve structures in an EM formed slide. Here the number of nodes that exist in the image describes how dense nerve fibers are in a particular region of the skin. Hildith proposes a network structural analysis algorithm for the automatic classification of chromosome spreads (type, relative size and orientation).

Dual stain kit for the microscopic gram classification of ocular infection bacteria

1993 ◽  
Vol 51 ◽  
pp. 248-249
Author(s):  
Jacob S. Hanker ◽  
Dale N. Holdren ◽  
Kenneth L. Cohen ◽  
Beverly L. Giammara
Keyword(s):  
Contact Lenses ◽  
Ocular Infection ◽  
Gram Negative Bacteria ◽  
Gram Negative ◽  
Gram Staining ◽  
Ocular Infections ◽  
Different Types ◽  
Culture Positive ◽  
Increased Susceptibility

Keratitis and conjunctivitis (infections of the cornea or conjunctiva) are ocular infections caused by various bacteria, fungi, viruses or parasites; bacteria, however, are usually prominent. Systemic conditions such as alcoholism, diabetes, debilitating disease, AIDS and immunosuppressive therapy can lead to increased susceptibility but trauma and contact lens use are very important factors. Gram-negative bacteria are most frequently cultured in these situations and Pseudomonas aeruginosa is most usually isolated from culture-positive ulcers of patients using contact lenses. Smears for staining can be obtained with a special swab or spatula and Gram staining frequently guides choice of a therapeutic rinse prior to the report of the culture results upon which specific antibiotic therapy is based. In some cases staining of the direct smear may be diagnostic in situations where the culture will not grow. In these cases different types of stains occasionally assist in guiding therapy.

Ultrastructural Classification of Pituary Adenomas

1990 ◽  
Vol 48 (3) ◽  
pp. 258-258
Author(s):  
S. Arumugam ◽  
Sarasa Bharati Arumugam
Keyword(s):  
Cytoplasmic Organelle ◽  
Null Cell ◽  
Cell Adenoma ◽  
Mixed Cell ◽  
Light Microscope Level ◽  
Null Cell Adenoma ◽  
Growth Hormone Cell ◽  
Ultrastructural Characteristics ◽  
Organelle Morphology

Adenoaas of the pituitary are no longer classified based on their tinctorial affinity to dyes. With the advent of the newer methods of sophisticated technology, it is now possible to classify. These depending upon the type of hormone secreted based either on histochemical techniques or on ultrastructural characteristics. The latter provides an insight into the cytoplasmic organelle morphology which offers a delightful feast to the eye as well.This paper presents the ultrastructural characters of the pituitary adenoma as seen in Madras. 171 adenomas (124 males and 47 females) were seen during 1972-1989, classified at the light microscope level as 159 chromophobe, 2 basophilic, 4 eosinophilic and 6 mixed adenomas.Ultrastructural examination showed that the sparsely granular prolactin cell adenoma is the commonest adenoma to be encountered closely followed by the growth hormone cell adenoma, null cell adenoma, the mixed cell adenoma and others.

Sign in / Sign up

Export Citation Format

Share Document