Artificial Intelligence via Competitive Learning and Image Analysis for Endometrial Malignancies

2019 ◽  
Vol 8 (4) ◽  
pp. 38-54 ◽  
Author(s):  
Abraham Pouliakis ◽  
Niki Margari ◽  
Effrosyni Karakitsou ◽  
George Valasoulis ◽  
Nektarios Koufopoulos ◽  
...  
Keyword(s):  
Artificial Intelligence ◽  
Image Analysis ◽  
System Performance ◽  
Malignant Disease ◽  
Competitive Learning ◽  
Image Analysis System ◽  
Morphometric Features ◽  
Analysis System ◽  
Histological Confirmation

Objective of this study is to investigate the potential of an artificial intelligence (AI) technique, based on competitive learning, for the discrimination of benign from malignant endometrial nuclei and lesions. For this purpose, 416 liquid-based cytological smears with histological confirmation were collected, each smear corresponded to one patient. From each smear was extracted nuclear morphometric features by the application of an image analysis system. Subsequently nuclei measurement from 50% of the cases were used to train the AI system to classify each individual nucleus as benign or malignant. The remaining measurement, from the unused 50% of the cases, were used for AI system performance evaluation. Based on the results of nucleus classification the patients were discriminated as having benign or malignant disease by a secondary subsystem specifically trained for this purpose. Based on the results it was conclude that AI based computerized systems have the potential for the classification of both endometrial nuclei and lesions.

Artificial Intelligence for the Identification of Endometrial Malignancies

2020 ◽  
pp. 266-279
Author(s):  
Abraham Pouliakis ◽  
Niki Margari ◽  
Effrosyni Karakitsou ◽  
Evangelia Alamanou ◽  
Nikolaos Koureas ◽  
...  
Keyword(s):  
Neural Network ◽  
Artificial Intelligence ◽  
Image Analysis ◽  
Vector Quantization ◽  
System Performance ◽  
Morphometric Characteristics ◽  
Cell Nuclei ◽  
Performance Indices ◽  
And Performance

Aim of this article is to investigate the potential of Artificial Intelligence (AI) in the discrimination between benign and malignant endometrial nuclei and lesions. For this purpose, 416 histologically confirmed liquid-based cytological smears were collected and morphometric characteristics of cell nuclei were measured via image analysis. Then, 50% of the cases were used to train an AI system, specifically a learning vector quantization (LVQ) neural network. As a result, cell nuclei were classified as benign or malignant. Data from the remaining 50% of the cases were used to evaluate the AI system performance. By nucleic classification, an algorithm for the classification of individual patients was constructed, and performance indices on patient classification were calculated. The sensitivity for the classification of nuclei was 77.95%, and the specificity was 73.93%. For the classification of individual patients, the sensitivity was 90.70% and the specificity 82.79%. These results indicate that an AI system can have an important role in endometrial lesions classification.

scholarly journals Karyometric Comparison of Splenic and Gastric Marginal Zone Lymphomas

2012 ◽  
Vol 35 (4) ◽  
pp. 297-303
Author(s):  
Magdalena Styczeń ◽  
Joanna Szpor ◽  
Sergiusz Demczuk ◽  
Krzysztof Okoń
Keyword(s):  
Neural Networks ◽  
Image Analysis ◽  
Statistical Analysis ◽  
Lymphoid Tissue ◽  
Marginal Zone ◽  
Lymphoid Tissues ◽  
Image Analysis System ◽  
Correct Classification ◽  
Analysis System

Background: Marginal zone lymphomas are indolent B-cell lymphomas associated with autoimmunity and chronic inflammation. The two most frequent variants are mucosa associated lymphoid tissues marginal zone lymphomas and splenic marginal zone lymphomas. The aim of the study was to determine if it is possible to classify splenic and gastric lymphomas according to karyometric features.Methods: The material consisted of 16 splenic and 14 gastric lymphomas. The measurements were done with the AnalySIS image analysis system. In each case at least 100 nuclei were selected, and 19 different geometric parameters were measured.Results: On statistical analysis, the nuclei of splenic and gastric lymphomas showed differences in most parameters, but significant overlap of the values was present. Neural networks were trained and used for classification of the data. By this method, the nuclei were properly classified with a sensitivity of 0.75 and specificity of 0.71. In addition, in all the cases the majority of the nuclei were properly classified, thus allowing correct classification of all the cases into “splenic” or “gastric”.Conclusion: These results support the view that mucosa-associated lymphoid tissue lymphomas and splenic marginal-zone lymphomas are separate entities.

Chemical Classification and Particle Sizing of Foundry Sands

1985 ◽  
Vol 43 ◽  
pp. 388-389
Author(s):  
D.S. DeMiglio
Keyword(s):  
Image Analysis ◽  
Energy Dispersive Spectrometer ◽  
Image Analysis System ◽  
Representative Sampling ◽  
Foundry Sands ◽  
Large Atomic Number ◽  
Backscattered Electron ◽  
Sand Particles ◽  
Analysis System ◽  
Reclamation System

Much progress has been made in recent years towards the development of closed-loop foundry sand reclamation systems. However, virtually all work to date has determined the effectiveness of these systems to remove surface clay and metal oxide scales by a qualitative inspection of a representative sampling of sand particles. In this investigation, particles from a series of foundry sands were sized and chemically classified by a Lemont image analysis system (which was interfaced with an SEM and an X-ray energy dispersive spectrometer) in order to statistically document the effectiveness of a reclamation system developed by The Pangborn Company - a subsidiary of SOHIO.The following samples were submitted: unreclaimed sand; calcined sand; calcined & mechanically scrubbed sand and unused sand. Prior to analysis, each sample was sprinkled onto a carbon mount and coated with an evaporated film of carbon. A backscattered electron photomicrograph of a field of scale-covered particles is shown in Figure 1. Due to a large atomic number difference between sand particles and the carbon mount, the backscattered electron signal was used for image analysis since it had a uniform contrast over the shape of each particle.

Sign in / Sign up

Export Citation Format

Share Document