Medical Technologies Journal

Editorial

Abdeldjalil Khelassi — Mon, 11 Dec 2023 00:00:00 +0000

Welcome to the fifth volume of the international peer reviewed journal: Medical Technologies Journal. We are proud to announce, in this editorial, the re-launching of publication on the journal. The journal receives all medical domains: trials and, synthesis;findings and innovations. After a period of silence due to several technical problemsprincipally the COVID19 crises, we publish, for our community, this issue. It contains fourscientific articles one review in ophthalmology and three in medical technologies.

Infectious Complications of Contact Lenses: A Review of the Literature

Zineb Tadj, Souad Taleb — Mon, 11 Dec 2023 00:00:00 +0000

The use of contact lenses is very common, they are prescribed for the correction of refractive errors that cannot be treated by glasses such as aphakia, keratoconus and strong anisometropia, or as alternatives to glasses. Contact lenses can cause serious complications that are not always easy for patients to manage [1]. Infectious complications in contact lens wearers are a real diagnostic and therapeutic emergency. These infections can be bacterial, amoebic (due to wearing contact lenses in swimming pools, rinsing lenses with tap water or saliva), fungal (due to wearing therapeutic, cosmetic or aphakic lenses, diabetes, alcoholism, immunosuppression, corticosteroids). Infectious complications require emergency treatment. The first line of treatment consists of removing the contact lens and sending it, along with the lens case and lens care product, to the laboratory along with corneal samples and appropriate medical treatment. Infectious keratitis related to contact lens wear is serious and can permanently affect the visual prognosis. The prevention requires a fitting under medical supervision and an awareness of the patients with potential risks of infection.

An Optimized Medical Image Watermarking Approach for E-Health Applications

Hadjer Abdi, Ismail Boukli Hacene — Mon, 11 Dec 2023 00:00:00 +0000

Background: In recent years, information and communication technologies have been widely used in the healthcare sector. This development enables E-Health applications to transmit medical data, as well as their sharing and remote access by healthcare professionals. However, due to their sensitivity, medical data in general, and medical images in particular, are vulnerable to a variety of illegitimate attacks. Therefore, suitable security and effective protection are necessary during transmission.

Method: In consideration of these challenges, we put forth a security system relying on digital watermarking with the aim of ensuring the integrity and authenticity of medical images. The proposed approach is based on Integer Wavelet Transform as an embedding algorithm; furthermore, Particles Swarm Optimization was employed to select the optimal scaling factor, which allows the system to be compatible with different medical imaging modalities.

Results: The experimental results demonstrate that the method provides a high imperceptibility and robustness for both secret watermark and watermarked images. In addition, the proposed scheme performs better for medical images compared with similar watermarking algorithms.

Conclusion: As it is suitable for a lossless-data application, IWT is the best choice for medical images integrity. Furthermore, using the PSO algorithm enables the algorithm to be compatible with different medical imaging modalities.

Classification of histological images of thyroid nodules based on a combination of Deep Features and Machine Learning

Linda Bellal, Meriem Saim, Amina Benahmed, Kamila Khemis — Mon, 11 Dec 2023 00:00:00 +0000

Background: Thyroid nodules are a prevalent worldwide disease with complex pathological types. They can be classified as either benign or malignant. This paper presents a tool for automatically classifying histological images of thyroid nodules, with a focus on papillary carcinoma and follicular adenoma.

Methods: In this work, two pre-trained Convolutional Neural Network (CNN) architectures, VGG16 and VGG19, are used to extract deep features. Then, a principal component analysis was used to reduce the dimensionality of the vectors. Then, three machine learning algorithms (Support Vector Machine, KNearest Neighbor, and Random Forest) were used for classification. These investigations were applied to our database collection,

Results: The proposed investigations have been applied to our private database collection with a total of 112 histological images. The highest results were obtained by the VGG16 transfer deep feature and the SVM classifier with an accuracy rate equal to 100%.

Categorization of Emotion Based on Causality

Farah Benayad, Djamel Bouchaffra, Faycal Ykhlef, Abdelkrim Allam — Mon, 11 Dec 2023 00:00:00 +0000

Background: Emotions come in all shapes and forms. Some of them can be external, visible, and clearly comprehensible, while others can seemingly be coming out of thin air. Knowing what causes an emotion is crucial for better therapy and mental health. Therefore, in this manuscript, we address the problem of emotions causality.

Methods: We propose a comparison of three traditional clustering models: Gaussian mixture model, HDBSCAN, and fuzzy c-means, to categorize each emotion described in the DEAP database. It contains over 1700 points, and has no prior label as to which type of stressor the subject’s emotion is generated from. This labelling task has been conducted by a psychiatrist.

Results: The fuzzy c-means yields the highest results, with an accuracy of 57.13%, followed by the Gaussian mixture model at 39.49% and the HDBSCAN method with only 18.86%. Another score computed is the mutual information score which shows how homogenous the clusters are for each model.

Conclusion: The data from DEAP is very heterogeneous and its density is stable,which may indicate that classification would be the better option, in terms of accuracy and homogeneity of the clusters.