Musa Ataş

Bio: Musa Ataş is an academic researcher from Siirt University. The author has contributed to research in topics: Hyperspectral imaging & Aflatoxin. The author has an hindex of 5, co-authored 20 publications receiving 132 citations. Previous affiliations of Musa Ataş include Middle East Technical University.

Hand Tremor Based Biometric Recognition Using Leap Motion Device

Abstract: In this paper, the applicability of hand tremor-based biometric recognition via leap motion device is investigated. The hypothesis is that the hand tremor is unique for humans and can be utilized as a biometric identification. In order to verify our hypothesis, spatiotemporal hand tremor signals are acquired from subjects. The objective is to establish a live and secure identification system to avoid mimic and cloning of password by attackers. Various feature extraction methods, including statistical, fast Fourier transform, discrete wavelet transform, and 1-D local binary pattern are used. For evaluating recognition performance, Naive Bayes and Multi-Layer Perceptron are utilized as linear-simple and nonlinear-complex classifiers, respectively. Since the conducted experiments produced promising results (above 95% of classification accuracy rate), it is considered that the proposed approach has the potential to be used as a new biometric identification manner in the field of security.

Image Steganalysis in Frequency Domain Using Co-Occurrence Matrix and Bpnn

Abstract: In the last two decades, steganalysis has become a fertile research area to minimize the security risks left behind by Misuse of data concealment in digital computer files. As the propagation of hidden writing increased, the need for the steganalysis emerged and grew to a large extent necessary to deter illicit secret communications. This paper introduces a steganalysis system to detect hidden information in images through using co-occurrence matrix, frequency domain transform, the first three moments, and back propagation neural network (BPNN). Four varieties of the system implemented. Firstly, the co-occurrence matrix calculated for the input image, which suspected to be a carrier of hidden secret information. Second, three levels of discrete wavelet transform (DWT) are applied resulting in 12 subbands. Then, those subbands along with the original image are transformed by discrete Fourier transform (DFT) or discrete cosine transform (DCT) to produce 13 subbands. After that, the first three moments are calculated resulting feature vector with 39 features. Finally, BPNN is used as a classifier to determine whether the image is containing hidden information or not. The system is tested with and without co-occurrence matrix, each of them once using DFT and another time using DCT. The results have shown that using co-occurrence matrix with DFT has the highest performance, which was 81.82% on the Hiding Ratio of 0.5 bit per pixel. This work demonstrates a good effect comparing to previous works.

A machine learning approach to dental fluorosis classification

Abstract: Fluoride in groundwater has been found to pose a severe public health threat in two villages (Karatas and Sarim) of western Sanliurfa in the southeastern Anatolia region of Turkey, where many cases of fluorosis, which detrimentally affects the teeth and bones, have been reported Analysis of fluoride in drinking water is usually accomplished using various chemical methods, but while these techniques produce accurate and reliable results, they are expensive, labor-intensive, and cumbersome In this study, a more cost-effective alternative, based on machine learning methods, is introduced In this case, artificial neural network (ANN), support vector machine (SVM), and Naive Bayes classifiers are utilized Furthermore, a novel feature selection and ranking method known as Normalized Weighted Voting Map (NWVM) is presented In Fisher discrimination power (FDP) scores, X-ray fluorescence (XRF) variables have higher discrimination power potential than X-Ray diffraction (XRD) attributes, the most salient feature being Zr (0464) and CaO (219993) from XRD and XRF, respectively When the XRD and XRF parameters are classified separately for the effect of NWVM ranking scores on the fluoride values and dental fluoride in groundwater, CaO, SiO2, MgO, Fe2O3, P2O5, and K2O (for XRF) and Quartz and Zr (for XRD) present a stronger effect In addition, when looking at the effects among themselves, the first order is the same XRF parameters and then the XRD parameters Experiments revealed that XRF constituents including CaO, SiO2, MgO, P2O5, and K2O have higher class discrimination power than the XRD variables

Aflatoxin contaminated chili pepper detection by hyperspectral imaging and machine learning

Abstract: Mycotoxins are toxic secondary metabolites produced by fungi. They have been demonstrated to cause various health problems in humans, including immunosuppression and cancer. A class of mycotoxins, aflatoxins, has been studied extensively because they have caused many deaths particularly in developing countries. Chili pepper is also prone to aflatoxin contamination during harvesting, production and storage periods. Chemical methods to detect aflatoxins are quite accurate but expensive and destructive in nature. Hyperspectral and multispectral imaging are becoming increasingly important for rapid and nondestructive testing for the presence of such contaminants. We propose a compact machine vision system based on hyperspectral imaging and machine learning for detection of aflatoxin contaminated chili peppers. We used the difference images of consecutive spectral bands along with individual band energies to classify chili peppers into aflatoxin contaminated and uncontaminated classes. Both UV and halogen illumination sources were used in the experiments. The significant bands that provide better discrimination were selected based on their neural network connection weights. Higher classification rates were achieved with fewer numbers of spectral bands. This selection scheme was compared with an information-theoretic approach and it demonstrated robust performance with higher classification accuracy.

Multimodal Emotion Recognition using Deep Learning

Abstract: New research into human-computer interaction seeks to consider the consumer's emotional status to provide a seamless human-computer interface. This would make it possible for people to survive and be used in widespread fields, including education and medicine. Multiple techniques can be defined through human feelings, including expressions, facial images, physiological signs, and neuroimaging strategies. This paper presents a review of emotional recognition of multimodal signals using deep learning and comparing their applications based on current studies. Multimodal affective computing systems are studied alongside unimodal solutions as they offer higher accuracy of classification. Accuracy varies according to the number of emotions observed, features extracted, classification system and database consistency. Numerous theories on the methodology of emotional detection and recent emotional science address the following topics. This would encourage studies to understand better physiological signals of the current state of the science and its emotional awareness problems.