Wyniki wyszukiwania - BazTech

1

Convolutional Neural Networks for C. Elegans Muscle Age Classification Using Only Self-learned Features

Czaplewski Bartosz, Dzwonkowski Mariusz, Panas Damian

Journal of Telecommunications and Information Technology

|

2022

|

nr 4

85--94

EN

Nematodes Caenorhabditis elegans (C. elegans) have been used as model organisms in a wide variety of biological studies, especially those intended to obtain a better understanding of aging and age-associated diseases. This paper focuses on automating the analysis of C. elegans imagery to classify the muscle age of nematodes based on the known and well established IICBU dataset. Unlike many modern classification methods, the proposed approach relies on deep learning techniques, specifically on convolutional neural networks (CNNs), to solve the problem and achieve high classification accuracy by focusing on non-handcrafted self-learned features. Various networks known from the ImageNet Large Scale Visual Recognition Challenge (ILSVRC) have been investigated and adapted for the purposes of the C. elegans muscle aging dataset by applying transfer learning and data augmentation techniques. The proposed approach of unfreezing different numbers of convolutional layers at the feature extraction stage and introducing different structures of newly trained fully connected layers at the classification stage, enable to better fine-tune the selected networks. The adjusted CNNs, as featured in this paper, have been compared with other state-of-art methods. In anti-aging drug research, the proposed CNNs would serve as a very fast and effective age determination method, thus leading to reductions in time and costs of laboratory research.

2

CBIR aided classification using extreme learning machine with probabilistic scaling in MRI brain image

Geetha A., Gomathi N.

Bio-Algorithms and Med-Systems

|

2020

|

Vol. 16, no. 2

art. no. 20190060

EN

An enormous number of magnetic resonance imaging (MRI) brain images were produced in hospitals and several MRI centers. To exploit the diagnosis in MRI brain image, “content-based image retrieval (CBIR)” system is accessed in the MRI brain image database. In this paper, a content-based MRI brain image retrieval system is presented, which is helpful in the medical field to seek a diagnosis in an MRI brain image that is similar to the example given. This paper consists of preprocessing, feature extraction, feature selection, similarity measure, and classification. In the preprocessing phase, the Wiener filter is used to remove the unwanted pixels from an MRI brain image. In the second phase, the features related to MRI brain image are extracted using characteristics of shape, margin, and density of the MRI. In the third stage, the features of MRI brain image were reduced using principal component analysis. CBIR classification is used in this method to gain effectual results. In the first stage, retrieval images are obtained using similarity measures using the similarity between the query image features and the derived trained image features. Finally, the classification stage is an extreme learning machine with probabilistic scaling used to classify the obtained retrieval output image and the query image. The result demonstrates that the proposed CBIR approach is robust and effectual compared with other latest work.

3

Thermal modelling and screening method for skin pathologies using active thermography

Strąkowska M., Strąkowski R., Strzelecki M., De Mey G., Więcek B.

Biocybernetics and Biomedical Engineering

|

2018

|

Vol. 38, no. 3

602--610

EN

This paper presents a novel screening approach of human skin pathologies using Active IR Thermography. The inputs of the proposed algorithm are the values of the physical parameters of the skin. Parameters are estimated based on dynamic thermographic measurements of human skin and the developed thermal model of the tissue. The calculations were based on the inverse thermal modelling. Classification was done using Support Vector Machine, Linear Discriminant Analysis and k-Nearest Neighbours classifiers. As an example, one presented the results of screening for psoriasis.

4

Use of two-dimensional matched filters for estimating a length of blood vessels newly created in angiogenesis process

Pająk R.

Opto-Electronics Review

|

2003

|

Vol. 11, No. 3

237-241

EN

The complex algorithm for estimating a length of blood vessels' newly created in angiogenesis process has been presented. A filtering method using two-dimensional matched filter is a fundamental process in the presented algorithm. For proper extraction of blood vessels' network more image processing techniques has been used including spatial low-pass filtering, binarisation, skeletonisation, and, developed by the author, algorithms for cleaning vessel network obtained in this process and for removing vessels false detected in the previous steps. The presented algorithm can be used for fully automatic vessels' length estimating in medical images and can be helpful for angiogenesis process's research.