Hanene Sahli

A new preprocessing parameter estimation based on geodesic active contour model for automatic vestibular neuritis diagnosis

The diagnostic of the vestibular neuritis (VN) presents many difficulties to traditional assessment methods This paper deals with a fully automatic VN diagnostic system based on nystagmus parameter estimation using a pupil detection algorithm. A geodesic active contour model is implemented to find an accurate segmentation region of the pupil. Hence, the novelty of the proposed algorithm is to speed up the standard segmentation by using a specific mask located on the region of interest. This allows a drastically computing time reduction and a great performance and accuracy of the obtained results. After using this fast segmentation algorithm, the obtained estimated parameters are represented in temporal and frequency settings. A useful principal component analysis (PCA) selection procedure is then applied to obtain a reduced number of estimated parameters which are used to train a multi neural network (MNN). Experimental results on 90 eye movement videos show the effectiveness and the accuracy of the proposed estimation algorithm versus previous work.

A novel automatic diagnostic approach based on nystagmus feature selection and neural network classification

Diagnosis of vertigo disorders presents many complications in the evaluation and treatment. In clinical practice, videonystagmography (VNG) tests are still an excellent bedside examination tool for vestibular disorder diagnosis. The parameters of different tests are used to get significant medical characterization of this disease. In this paper, we propose an approach to develop the assessment of vertigo symptom by the selection of the most pertinent VNG parameters using Fisher Linear Discriminant analysis. Therefore, a multilayer neural network (MNN) classifier is applied for automatic VNG dataset analysis based upon the fundamental measurements of normal and affected patients by vestibular disorder. The experimental results prove that the proposed approach is very interesting and helpful for an accurate diagnostic of this disease.

A new approach based on the serological tests for the diagnosis of tuberculosis in cattle

The diagnosis of bovine tuberculosis (TB) is still a challenge for a better control of the disease. Here we report the use of a simple ELISA test combined with a multilayer neuronal network analyzing method in order to diagnose TB in cattle from the north part of Tunisia. A panel of Mycobacterium bovis (Mbv)-specific recombinant proteins along with crud extracts were used to coat the 96 wells plates, namely the recombinant 10 kDa culture filtrate antigen (CFP-10), the 6 kD Early Secretory Antigenic Target (ESAT-6), the recombinant Esat-6/CFP-10 heterodimer along with the crud BCG proteins and Tuberculin purified protein derivative (PPD). In the current article, a new approach is described to compare their characterization degree to select the most discriminative antigens. The classification of subjects into two groups: TB+ and TB-subjects was affected by an artificial multilayer neural network and the statistical study to estimate the diagnosis of bovine tuberculosis and construct an optimal partition of the results. This method was applied on the serological tests of a set of cattles. The results are encouraging.

An advanced intelligent ELISA test for bovine tuberculosis diagnosis

Abstract The control of bovine tuberculosis (bTB) relies first on an optimal diagnosis of the disease. Several tests have been implemented for bTB detection which are generally complex, slow of use and relatively expensive especially in poor countries. A simple rapid, cost effective and efficient automated method for bTB assessment is still needed. Here, we propose a combination of the simple Enzyme Linked Immuno Sorbent Assay (ELISA) test with either the artificial neural network (ANN) analyzing method to effectively diagnose TB in cattle. The proposed method has been experimented on 30 bTB+ and 43 bTB- subjects in the north part of Tunisia, as assessed by the intra dermal reaction test (IDR). The obtained results have reached a 94% of accuracy when applying the ANN. Moreover, the proposed methodology enabled us to reduce the number of the used pathogens-derived antigens to three instead of the standard five antigens-based ELISA. Compared to previous works, the proposed expert system seems to be promising and may prove helpful for the veterinary diagnosis of tuberculosis.

A comparative study applied to dynamic textures segmentation

As textures in motion, dynamic textures (DT) are video sequences that are spatially and temporally repetitive. Since, the segmentation of this type of textures represents many difficulties; this paper proposes a comparative study between neural, fuzzy and statistical methods for DT segmentation. In order to bring the best results for DT classification, the proposed method is based on an extraction of two major areas of sequence into two categories: static zone (SZ) and dynamic zone (DZ). Then, for the same data, we applied our methods to compare between dynamic textures DT and static textures (ST). By extracting the best method and thanks to the Multilayer artificial Neural Network (MNN) segmentation, 82% is achieved as classification accuracy. Finally, the experimental results show that the three used scenes are successfully applied to DT.

Automated detection of current fetal head in ultrasound sequences

Accurate diagnostic and prognostic of fetus detects is an important challenge based on fetal head formation to supply much critical information that requires more attention in evaluating the abnormal heads. One of the fundamental problems currently faced, is how to limit the low signal to noise ratio with respect to the complexity of small fetal head ultrasound images dimension. This paper deals with a fully automatic detection system of subsequent fetal head composition from ultrasound images. In the preprocessing task, two filters have been used for speckle noise reducing. Using the Hough transform technique, fetal head structure detection is achieved, giving 97% as segmentation accuracy. Experimental results are analyzed using five ultrasound sequences that illustrate the effectiveness and the accuracy of the proposed method for a factual diagnostic of fetal heads.

New approach based on fuzzy classification of the serological tests “ELISA” for the diagnosis of cattle tuberculosis

Several antigens have been produced and/or secreted from the mycobacterium bovis. The agent of bovine tuberculosis (TB) is used to detect this disease in cattle through the serological ELISA test (a, b, c, d and e). In this work, we propose a novel approach to improve the diagnosis bovine tuberculosis. In order to select the antigens with top priority, the proposed methodology is based on a comparison of their power of characterization. Experimental results are analyzed using two categories: sick (TB +) and not sick (TB -). By extracting some original features and thanks to the unsupervised Fuzzy C-Means (FCM) classification, 89% is achieved as classification accuracy. Compared to previous works, the proposed expert system is very promising and helpful for the veterinary diagnosis of tuberculosis.

A new approach based on the serological tests and the Delayed Hyper Sensitivity Tests for the diagnosis of canine leishmaniasis

Several responses in the form of serological tests ELISA (Enzyme Linked Immuno Sorbent Assay), IIF (Indirect Immuno Fluoresence) and DHST (Delayed Hyper Sensitivity Tests) can be used to detect leishmania parasite infection in dogs. In this paper, we propose a new method to select the most discriminative tests based on determinant criterion. So, the diagnosis of canine leishmaniasis (CanL) can be improved by reducing the number of features. Moreover, an artificial neural networks (the Multilayer Preceptron neural network) is applied to classify subjects into two groups: positive (sick) and negative (not sick). The correlation between the physical and the pathological state of subjects is specified with multiple attempts. These methods are obtained with considering chain of experiences that allow for fairly reliable and highly effective results which enable us to develop an efficient way to estimate the diagnosis of this disease. After many experiments, we notice that the best combination of the three studied tests is the DHST and IIF tests.