International Journal of Interactive Mobile Technologies(iJIM) – eISSN: 1865-7923 – Vol. 16 No. 08 (2022) Paper—Face Patterns Analysis and Recognition System Based on Quantum Neural Network QNN Face Patterns Analysis and Recognition System Based on Quantum Neural Network QNN https://doi.org/10.3991/ijim.v16i08.30107 Haider TH. Salim ALRikabi1(*), Ibtisam A. Aljazaery2, Jaafar Sadiq Qateef 1, Abdul Hadi M. Alaidi1, Roa’a M. Al_airaji2 1Wasit University, Wasit, Iraq 2University of Babylon, Babylon, Iraq hdhiyab@uowasit.edu.iq Abstract—The past few years have witnessed a huge increase in the appli- cation of facial recognition, detection, and analysis technology. However, face recognition systems remain the most popular among the general public. The facial recognition system can detect the presence of a face when exposed to one. The accuracy and fairness that can be derived from such systems necessitate their use, because humans, particularly security personnel can be tired and target the wrong person as a suspect. However, artificial intelligence systems that are prop- erly trained are capable of efficiently identifying and classifying faces without errors. In this work, the use of Matlab language was employed in building a software system that is capable of recognizing and differentiating different face patterns. The proposed system is equipped with a camera that serves as the prac- tical aspect of the software that captures different shots that are sent to the theo- retical part of a special program that is designed to recognize faces by comparing them with a database stored within the program. The practical part of the work involved the use of Quantum Neural Network. In this work, the training dataset is made up of features vectors that were obtained from a well-known set of face images of different people. Here, Principle Component Analysis (PCA) was used for the extraction of feature vectors from images and then prepared for the next training step. The experimental results revealed that efficient face recognition can be achieved through the use of well-trained Quantum Network. Keywords—artificial intelligence, facial recognition, feature extraction, QNN, PCA 1 Introduction Academics and industry professionals are increasingly becoming interested in the accuracy of computerized identification systems, due to the increasing security chal- lenges around the world [1]. This interest has resulted in a drastic increase in the number of security applications and systems. A wide variety of algorithms ranging from simple to complex ones have been developed to this end [2, 3]. However, the main question here is “how accurate is the facial recognition algorithm for security applications?” 34 http://www.i-jim.org https://doi.org/10.3991/ijim.v16i08.30107 mailto:hdhiyab@uowasit.edu.iq Paper—Face Patterns Analysis and Recognition System Based on Quantum Neural Network QNN Over the years, a wide range of commercial products has been developed by the Academic Computer Science Resellers [4, 5]. These products have improved the effi- ciency of automated identification algorithms in different fields where faces recogni- tion tasks are required. Consequently, the field has advanced with a view of producing systems of high resolution that can effectively perform the task of face identification. Hence, this field appeared to be developed to get a very high-resolution system to iden- tify the faces, which in turn is a final number of applications that begin with children’s education programs and end with global security systems [6, 7]. Below is a summary of the steps involved in facial recognition systems: 1. ACQUIRING: Obtain image by capturing. 2. DETECTING: face image is extracted from the entire image. 3. ALIGNMENT: at this stage, the image is subjected to the processes of alignment and image sclerosis, whereby, the facial angle is adjusted to the camera angle. 4. EXTRACTING: the most significant features of the acquired image are extracted at this stage. 5. MATCHING: at this stage, the required image and photo store are matched. 6. REPORTING: The stage of issuing a report closer to the image or dissimilar. Out of the different measures of biometrics, the most natural is facial recognition. Naturally, this can be regarded as reasonable, because humans are typically able to recognize themselves just by looking at each other’s, rather than through irises and thumbprints. More than half of the world has been estimated to be influenced by facial recognition technology in one way or the other [8–10]. The automated facial recogni- tion system works by identifying an individual based on a given picture of that person which was previously stored in the memory of the system [11, 12]. The memory of the facial recognition system is created and extracted from a training dataset. The feature vector refers to any collection chosen from the matrices of the original image. More so, the features vectors are representative of the most significant and critical values in the original images, and hence the reduction of the images’ size to vectors that summarize the images. Therefore, the recognition is performing the task of identifying the fea- ture vector which is very similar or closest training set to the feature vector extracted from the image presented for testing (i.e., the image whose identity must be known using the recognition system. A person’s identity can be known and differentiated by inputting the person’s image into the system; the image that is inputted into the system for identification is referred to as the test image [13–16]. One of the most efficient statistical techniques that have been employed in the field of image recognition and image compression, is the PCA algorithm, whose role is centered on the reduction of large dimensions of data space to smaller dimensions spaces. Under normal con- ditions, the new spaces are feature spaces, containing the essential and most relevant features of the data within their original spaces. Thus, the size of the data is reduced and made portable through the reduction of dimensions. The reduction of the dimensions is much appreciated at a later stage. The reason why reducing the dimensions is important is that the face recognition systems normally encounter problems when dealing with spaces of large dimensions such as images. There are several ways through which this issue can be addressed; extant data can be matched and transferred to data within a iJIM ‒ Vol. 16, No. 08, 2022 35 Paper—Face Patterns Analysis and Recognition System Based on Quantum Neural Network QNN less dimensional [17–19]. Therefore, the process of reducing the dimensionality begins from the original space of large dimensions and then moves to the new space containing smaller dimensions. Assuming that the following vector in a space of N dimension: x x x x xN T� ��� ��1 2 3, , , .., (1) After the reduction of dimensions, it will become another vector in another space of K dimension: y y y y yK T� ��� ��1 2 3, , , .., (2) Where K < N The PCA algorithm is mainly used for the reduction of the data’s dimensionality, while as much as possible relevant information is retained within the original data [20]. The process is almost the same as keeping as many variations and modifications as pos- sible in the original data. The data within the space of higher dimensionality is mapped out by the PCA to the matching information within the lower-dimensional subspace; the PCA achieves this through a linear transformation as shown below: y Tx= � (3) Where T denotes the transformation in which the value of x y− is smaller. The areas of quantum information and quantum computation have witnessed rapid growth, thereby increasing the potentials of quantum neural networks (QNN’s. The actualization of QNN has been stimulated by the drastic advancement of quantum computer hardware. Given the quantum properties possessed by the QNN, it is com- putationally efficient and has greater storage capacity as compared to its counterparts. The areas of disease prediction, image processing, speech recognition, as well as other fields have seen extensive use of QNN. In academia, there are several definitions of QNN, but there is no consensus on a specific definition. However, the most acceptable one defines QNN as a computing paradigm that is a combination of neuroscience and quantum computing [21]. Particularly, quantum theory and neural network model are connected by QNN; QNN achieves this by considering the similarity between the basic quantum computing, two-level qubits, and the active/idle states within the complex process of signal transmission in the nerve cells. It is clear that the system evolves based on quantum effects and it is in conformity to the principle theories of quantum mechanics. Nevertheless, the majority of the QNN models proposed are centered on the aspect of mathematical computations. Such models have failed to address impending issues like the ambiguous physical feasibility of QNN. Also, some of them do not fol- low the changes that have occurred in the area of quantum effects and are not designed with the characteristics of neural network computing. For this reason, the actual QNNs have not been actualized in real life. Based on a historical development view, there are approximately two phases through which QNN has passed; early-stage and near-term quantum processor stage. previously, it was impossible to implement QNN on quantum 36 http://www.i-jim.org Paper—Face Patterns Analysis and Recognition System Based on Quantum Neural Network QNN computers because of the conditions of the hardware. At that time the proposals of several models were made based on the physical process associated with quantum com- puting, without any description of particular quantum processor structures like quantum circuits and qubits. The classic example of these kinds of QNNs includes QNN based on dissipative quantum gates, QNN based on multiple world views, QNN analogous to CNN, QNN based on interactive quantum dots, etc. In comparison to results from early studies, the QNN proposed in recent times is broader in meaning. Usually, the concept of QNN is used when referring to a computational model that is characterized by a networked structure and parameters that can be trained, and are implemented through a quantum system and quantum circuit. Furthermore, recent studies have addressed the problem of the model’s physical feasibility, as the emphasis is now placed on that aspect of the model. There are several advantages that the QNN offers. Nevertheless, at the current stage of QNN development, the inconsistency between unitary dynamics in quantum computing and dissipative dynamics in neural computing is an area that is begging for a more comprehensive study. In addition, the present QNN is limited by the fact that it can only be trained for larger samples at low latitudes, and the accuracy of prediction and generalization of performance remains a problem yet to be solved. More so, the parameter space of the QNN exponential level experiences the easy formation barren plateau phenomenon. 2 Related work For the first time in 1980, the concept of quantum computation was introduced by Benioff [22], while the quantum neural computing concept was introduced in 1995 by Kak [4]. Subsequently, a study was conducted by Kak [23] who investigated the concept of quantum neural computing from the perspective of different new directions in neural network research. Also, in 1995, in a study carried out by Menneer and Narayanan [5] quantum-inspired neural network (QINN) was proposed. In another study, Perus [6] presented a discussion on the absorption comparability between quantum paral- lelism and neural networks. In [24], the use of optical interference was employed in proposing a hypothetical model. To provide in-depth insight into the quantum neural network, the authors [25, 26] presented a comprehensive description as well as a sys- tematic review of the quantum neural network. In the work of Ventura and Martinez [9] which was carried out in the year 2000, the quantum implementation of the associative memory model was proposed. Similarly, Behrman and his associates [27] introduced the idea of physically implementing the QNN as a collection of dots. By means of simulation and experiments, Narayanan and Menneer [28] demonstrated that quantum ANN is not only more. A quantum perception model was introduced by Altaisky [29], who named the model “network with a teacher”. Presently, the majority of the models of QNN proposed are self-organized network that does not require human interven- tion. Also, in the work done by Gupta and Zia [30], a new computational model was defined; the authors named it QNN based on the concept of Deutsch’s model of the quantum computational network. In 2003, Kouda et al. [31] proposed Qubit neural network. Karayiannis et al. proposed a model based on multi-level transfer function iJIM ‒ Vol. 16, No. 08, 2022 37 Paper—Face Patterns Analysis and Recognition System Based on Quantum Neural Network QNN Quantum Neural Networks [32]. The multilevel transfer function is adopted by the transfer function of neural cell within hidden layer since the feature of fuzziness is possessed by the network. The model which is based on multi-level transfer function QNN has proven to be of great importance in solving indetermination and crossed pattern problem. In the study carried out in [33], the prediction of weather has been successfully carried out using multi-level transfer function QNN. Similarly, [34, 35] have also used it for voice recognition and diagnosis of disease, etc. QNN is also known as a feed-forward neural network with multiple layers. With the QNN, indeterminate data can be effectively classified [36–38]. The transfer function of the quantum neural network is expressed as a linear superposition of the multi-sigmoid function. Hence it is referred to as a multi-level transfer function [39]. A deep neural network consisting of nine layers was used in the model, which was afterward tested. The model achieved an accuracy rate of 97.35%. The use of three databases was employed in training the model, and the databases include, Youtube Faces (YTF), Labelled Faces in the Wild (LFW), and Social Face Classification (SFC) [28]. The authors in [40, 41] performed the tasks of extracting pattern recognition through the use of an algorithm referred to as Local Vector Projection Classification (LVPC). Six face databases (UMIST, AR, ORL, Yale B, FERET, and Yale). An average recognition rate of 89.48% was achieved by the system [42]. In another study, the authors tested a model designed based on the dictionary learning method. In their work, the model learned features by using Bilinear Discriminative Dictionary Learning (BDDL) [43], while the Collaborative representa- tion-based classification (CRC) was used for image classification in the model. In 2015, the authors in [40] enhanced the CNN model’s performance to 0.8763 by combining a fusion network with Joint Bayesian. Nevertheless, the CNN layers are dependent on the size of the data, and as such, there are 2 models used for small size and medium sizes. In other to minimize the training time, the GPU was used in operating the model. The model was tested using public database LFW (Labeled Faces in the Wild) which consists of 13,233 images [44]. It was observed that the extant commercial systems are faced with challenges whenever they are faced with real life tasks. Many authors have made efforts to solve this problem using many software such as Convolutional Neural Networks which has been used for estimation of age, gender, and emotion. The dataset was created by collecting 4 million images from 40,000 people. Nevertheless, human annotators have labelled the images with parts of emotion and gender labels. The proposed software, achieved age classification accuracy of 96.2%, while the score for mean absolute error was 5.76 when the images were annotated, and an accuracy rate of 61.3% when the data was not annotated [45]. 2.1 Normalization of transforming image data Several factors influence the extraction of features, including offset of data and peak-to-peak magnitudes. These occur as a result of the surrounding physiologi- cal conditions, artifacts, psychological state, and the parameters of the measurement system. To address the issue, the process of normalization is applied so that the effects of undesirable offset and parameters can be minimized. The process of normalizing transformed image data involves the following steps: 38 http://www.i-jim.org Paper—Face Patterns Analysis and Recognition System Based on Quantum Neural Network QNN 1. Set transformed data’s mean value to zero value. 2. Subtract the mean value from the original image data. 3. The mean value of the produced image data after subtraction should be zero or close to zero. 3 Proposing work outline Fig. 1. Block diagram of proposed work 3.1 Project requirements Two requirements were needed for the completion of this research project. The first is an image that was obtained using cameras. The captured images were saved and obtained from the LFW Face Database. The second requirement involved the construc- tion of special programs through the use of QNN technology as well as Matlab lan- guage for the recognition of faces. The software achieves this by sending the images after they have been processed and their most relevant details extracted to the program for the analysis and comparison between the database stored to analyze and compare them with the database stored within the program. iJIM ‒ Vol. 16, No. 08, 2022 39 Paper—Face Patterns Analysis and Recognition System Based on Quantum Neural Network QNN Fig. 2. LFW face database used in this work 3.2 Implementation The programs used in this study were built and implemented using the Matlab lan- guage. The PCA was used to extract the critical features from the image during the preprocessing, while the training set was used to train the quantum neural network; it was trained to differentiate faces using the database with which it was fed. Lastly, the testing set was used in testing the performance of the model, and the corresponding score was recorded. 40 http://www.i-jim.org Paper—Face Patterns Analysis and Recognition System Based on Quantum Neural Network QNN Fig. 3. Structure of QNN for face recognition 3.3 Results and discussion The QNN’s ability to recognize images was examined using the faces of five people. The results obtained have been compared using statistical methods. The performance of the model was evaluated based on the following parameters: Sensitivity (SE), positive predictivity (PP), and Total Classification Accuracy (TCA). The parameters are defined as following [46–48]: Sensitivity (SE) = TP TP FN i i i+ (4) Positive Predictivity (PP) = TP TP FP i i i+ (5) TCA = TP T i rj� � 1 5 (6) iJIM ‒ Vol. 16, No. 08, 2022 41 Paper—Face Patterns Analysis and Recognition System Based on Quantum Neural Network QNN Where (TPi) is the number of images i that have been correctly recognized. (FNi) is the number of miss recognition of image i. (FPi) is the number of correctly recognized images as another person of image i. (Tr) is the number of all images in the testing set. Table 1. Recognition results obtained using QNN and PCA TP FN FP Se PP Person 1 4 1 0 0.8 1.0 Person 2 5 0 0 1.0 1.0 Person 3 5 0 0 1.0 1.0 Person 4 4 0 1 1.0 0.8 Person 5 4 0 1 1.0 0.8 TCA 89% The effect of the number of epochs on the training phase of the network, is one of the results that were obtained, because there is a certain level of balance between the error rate emanating from the correct non-recognition and the number of those epochs, as shown in Figures 4, 5, 6. Fig. 4. Error value versus number of iterations (150 epochs) 42 http://www.i-jim.org Paper—Face Patterns Analysis and Recognition System Based on Quantum Neural Network QNN Fig. 5. Error value versus number of iterations (200 epochs) Fig. 6. Error value versus number of iterations (250 epochs) iJIM ‒ Vol. 16, No. 08, 2022 43 Paper—Face Patterns Analysis and Recognition System Based on Quantum Neural Network QNN 4 Conclusions This article presents a system that was specifically designed for the recognition of face patterns through the use of a quantum neural network as well as one component transformed by the PCA so that the most significant facial features can be extracted. Based on the results presented in the tables and figures, the quantum network is capa- ble of achieving correct recognition at a high percentage, even though, the database was not large enough to make generalizations. This is indicative of how efficient the quantum network is as well as the training accuracy. The PCA has demonstrated supe- rior performance in terms of the extraction of the most significant features of the face pattern so that a correct dataset can be built at the training phase. It is noteworthy that the training time represented by the number of epochs is proportional to the error rate which may cause the discrimination process. 5 References [1] N. J. Ogbuke, Y. Y. Yusuf, K. Dharma, and B. Mercangoz, “Big data supply chain analytics: ethical, privacy and security challenges posed to business, industries and society,” Produc- tion Planning Control, vol. 33, no. 2–3, pp. 123–137, 2022. https://doi.org/10.1080/095372 87.2020.1810764 [2] S. Abadal, A. Jain, R. Guirado, J. López-Alonso, and E. Alarcón, “Computing graph neural networks: a survey from algorithms to accelerators,” ACM Computing Surveys, vol. 54, no. 9, pp. 1–38, 2021. https://doi.org/10.1145/3477141 [3] H. Salim, “Enhanced data security of communication system using combined encryption and steganography,” International Journal of Interactive Mobile Technologies, vol. 15, no. 16, pp. 144–157, 2021. https://doi.org/10.3991/ijim.v15i16.24557 [4] P. Kaur, K. Krishan, S. K. Sharma, and T. Kanchan, “Facial-recognition algorithms: a lit- erature review,” Medicine, Science the Law, vol. 60, no. 2, pp. 131–139, 2020. https://doi. org/10.1177/0025802419893168 [5] S. Khan, M. H. Javed, E. Ahmed, S. A. Shah, and S. U. Ali, “Facial recognition using convo- lutional neural networks and implementation on smart glasses,” in 2019 International Con- ference on Information Science and Communication Technology (ICISCT), 2019, pp. 1–6: IEEE. https://doi.org/10.1109/CISCT.2019.8777442 [6] K. Basu, and Z. Tzannatos, “The global child labor problem: what do we know and what can we do?” The world bank economic review, vol. 17, no. 2, pp. 147–173, 2003. https://doi. org/10.1093/wber/lhg021 [7] A. H. Alaidi, C. S. Der, and Y. W. J. i. Leong, “Systematic review of enhancement of arti- ficial bee colony algorithm using ant colony pheromone,” International Journal of Interac- tive Mobile Technologies (iJIM), vol. 15, no. 16, p. 173, 2021. https://doi.org/10.3991/ijim. v15i16.24171 [8] Z. Ming, M. Visani, M. M. Luqman, and J.-C. Burie, “A survey on anti-spoofing methods for facial recognition with rgb cameras of generic consumer devices,” Journal of Imaging, vol. 6, no. 12, p. 139, 2020. https://doi.org/10.3390/jimaging6120139 [9] I. A. Aljazaery, and H. ALRikabi, “Encryption of color image based on DNA strand and exponential factor,” International journal of online and biomedical engineering (iJOE), vol. 18, no. 3, 2022. https://doi.org/10.3991/ijoe.v18i03.28021 44 http://www.i-jim.org https://doi.org/10.1080/09537287.2020.1810764 https://doi.org/10.1080/09537287.2020.1810764 https://doi.org/10.1145/3477141 https://doi.org/10.3991/ijim.v15i16.24557 https://doi.org/10.1177/0025802419893168 https://doi.org/10.1177/0025802419893168 https://doi.org/10.1109/CISCT.2019.8777442 https://doi.org/10.1093/wber/lhg021 https://doi.org/10.1093/wber/lhg021 https://doi.org/10.3991/ijim.v15i16.24171 https://doi.org/10.3991/ijim.v15i16.24171 https://doi.org/10.3390/jimaging6120139 https://doi.org/10.3991/ijoe.v18i03.28021 Paper—Face Patterns Analysis and Recognition System Based on Quantum Neural Network QNN [10] M. Roa’a, I. A. Aljazaery, S. K. Al_Dulaimi, H. T. Salim, “Generation of high dynamic range for enhancing the panorama environment,” Bulletin of Electrical Engineering, vol. 10, no. 1, 2021. https://doi.org/10.11591/eei.v10i1.2362 [11] R. J. Baron, “Mechanisms of human facial recognition,” International Journal of Man-Machine Studies, vol. 15, no. 2, pp. 137–178, 1981. https://doi.org/10.1016/ S0020-7373(81)80001-6 [12] P. B. Balla and K. Jadhao, “IoT based facial recognition security system,” in 2018 interna- tional conference on smart city and emerging technology (ICSCET), 2018, pp. 1–4: IEEE. https://doi.org/10.1109/ICSCET.2018.8537344 [13] M. Mehra, V. Sahai, P. Chowdhury, and E. Dsouza, “Home security system using iot and aws cloud services,” in 2019 International Conference on Advances in Computing, Communication and Control (ICAC3), 2019, pp. 1–6: IEEE. https://doi.org/10.1109/ ICAC347590.2019.9089839 [14] I. A. Aljazaery, A. A. Ali, and H. M. Abdulridha, “Classification of electroencephalograph (EEG) signals using quantum neural network,” Signal Processing: An International Journal, vol. 4, no. 6, p. 329, 2011. [15] H. Tuama, and N. Alseelawi, “A novel method of multimodal medical image fusion based on hybrid approach of NSCT and DTCWT,” International journal of online and biomedical engineering, vol. 18, no. 3, 2022. https://doi.org/10.3991/ijoe.v18i03.28011 [16] R. M. ALairaji, and I. Aljazaery, “Abnormal behavior detection of students in the exam- ination hall from surveillance videos,” in Advanced Computational Paradigms and Hybrid Intelligent Computing, vol. 1373: Springer Singapore, 2022, pp. 113–125. https://doi. org/10.1007/978-981-16-4369-9_12 [17] R. A. Azeez, M. K. Abdul-Hussein, and M. S. Mahdi, “Design a system for an approved video copyright over cloud based on biometric iris and random walk generator using water- mark technique,” Periodicals of Engineering Natural Sciences, vol. 10, no. 1, pp. 178–187, 2021. ] [18] A. McCallum, K. Nigam, and L. H. Ungar, “Efficient clustering of high-dimensional data sets with application to reference matching,” in Proceedings of the sixth ACM SIGKDD international conference on Knowledge discovery and data mining, 2000, pp. 169–178. https://doi.org/10.1145/347090.347123 [19] O. H. Yahya, H. T. S. Alrikabi, and I. A. Aljazaery, “Reducing the data rate in internet of things applications by using wireless sensor network,” International journal of online and biomedical engineering, Article vol. 16, no. 3, pp. 107–116, 2020. https://doi.org/10.3991/ ijoe.v16i03.13021 [20] F. Gao, T. Tian, T. Yao, and Q. Zhang, “Human gait recognition based on multiple feature combination and parameter optimization algorithms,” Computational Intelligence Neurosci- ence, vol. 2021, 2021. https://doi.org/10.1155/2021/6693206 [21] Z. Sabir, M. A. Z. Raja, J. L. Guirao, and M. Shoaib, “Integrated intelligent computing with neuro-swarming solver for multi-singular fourth-order nonlinear Emden–Fowler equation,” Computational Applied Mathematics, vol. 39, no. 4, pp. 1–18, 2020. https://doi.org/10.1007/ s40314-020-01330-4 [22] P. Benioff, “The computer as a physical system: A microscopic quantum mechanical Hamiltonian model of computers as represented by Turing machines,” Journal of statistical physics, vol. 22, no. 5, pp. 563–591, 1980. https://doi.org/10.1007/BF01011339 [23] S. C. Kak, “Quantum neural computing,” Advances in imaging electron physics, vol. 94, pp. 259–313, 1995. https://doi.org/10.1016/S1076-5670(08)70147-2 iJIM ‒ Vol. 16, No. 08, 2022 45 https://doi.org/10.11591/eei.v10i1.2362 https://doi.org/10.1016/S0020-7373(81)80001-6 https://doi.org/10.1016/S0020-7373(81)80001-6 https://doi.org/10.1109/ICSCET.2018.8537344 https://doi.org/10.1109/ICAC347590.2019.9089839 https://doi.org/10.1109/ICAC347590.2019.9089839 https://doi.org/10.3991/ijoe.v18i03.28011 https://doi.org/10.1007/978-981-16-4369-9_12 https://doi.org/10.1007/978-981-16-4369-9_12 https://doi.org/10.1145/347090.347123 https://doi.org/10.3991/ijoe.v16i03.13021 https://doi.org/10.3991/ijoe.v16i03.13021 https://doi.org/10.1155/2021/6693206 https://doi.org/10.1007/s40314-020-01330-4 https://doi.org/10.1007/s40314-020-01330-4 https://doi.org/10.1007/BF01011339 https://doi.org/10.1016/S1076-5670(08)70147-2 Paper—Face Patterns Analysis and Recognition System Based on Quantum Neural Network QNN [24] I. Bretherton, “Communication patterns, internal working models, and the intergener- ational transmission of attachment relationships,” Infant mental health Journal, vol. 11, no. 3, pp. 237–252, 1990. https://doi.org/10.1002/1097-0355(199023)11:3<237::AI D-IMHJ2280110306>3.0.CO;2-X [25] B. Ricks and D. Ventura, “Training a quantum neural network,” Advances in neural infor- mation processing systems, vol. 16, 2003. [26] A. Khamparia and K. M. Singh, “A systematic review on deep learning architectures and applications,” Expert Systems, vol. 36, no. 3, p. e12400, 2019. https://doi.org/10.1111/ exsy.12400 [27] E. C. Behrman, J. Niemel, J. E. Steck, and S. R. Skinner, “A quantum dot neural network,” in Proceedings of the 4th Workshop on Physics of Computation, 1996, pp. 22–24. [28] A. Narayanan and T. Menneer, “Quantum artificial neural network architectures and compo- nents,” Information Sciences, vol. 128, no. 3–4, pp. 231–255, 2000. https://doi.org/10.1016/ S0020-0255(00)00055-4 [29] M. Altaisky, “Quantum neural network,” arXiv preprint quant-ph/0107012, 2001. [30] S. Gupta and R. Zia, “Quantum neural networks,” Journal of Computer System Sciences, vol. 63, no. 3, pp. 355–383, 2001. https://doi.org/10.1006/jcss.2001.1769 [31] N. Kouda, N. Matsui, H. Nishimura, and F. Peper, “Qubit neural network and its efficiency,” in International conference on knowledge-based and intelligent information and engineer- ing systems, 2003, pp. 304–310: Springer. https://doi.org/10.1007/978-3-540-45226-3_42 [32] G. Purushothaman and N. B. Karayiannis, “Quantum neural networks (QNNs): inherently fuzzy feedforward neural networks,” in Proceedings of International Conference on Neural Networks (ICNN‘96), 1996, vol. 2, pp. 1085–1090: IEEE. [33] D. Zhu and R. Wu, “A multi-layer quantum neural networks recognition system for hand- written digital recognition,” in Third International Conference on Natural Computation (ICNC 2007), 2007, vol. 1, pp. 718–722: IEEE. https://doi.org/10.1109/ICNC.2007.70 [34] G. Muhammad, M. Alsulaiman, A. Mahmood, and Z. Ali, “Automatic voice disorder clas- sification using vowel formants,” in 2011 IEEE international conference on multimedia and expo, 2011, pp. 1–6: IEEE. https://doi.org/10.1109/ICME.2011.6012187 [35] K. López-de-Ipiña, J.-B. Alonso, C. M. Travieso, J. Solé-Casals, H. Egiraun, M. Faundez-Zanuy, A. Ezeiza, N. Barroso, M. Ecay-Torres, and P. Martinez-Lage, “On the selection of non-invasive methods based on speech analysis oriented to automatic Alzheimer disease diagnosis,” Sensors, vol. 13, no. 5, pp. 6730–6745, 2013. https://doi.org/10.3390/ s130506730 [36] S. M. Taha and Z. Taha, “EEG signals classification based on autoregressive and inherently quantum recurrent neural network,” International Journal of Computer Applications in Tech- nology, vol. 58, no. 4, pp. 340–351, 2018. https://doi.org/10.1504/IJCAT.2018.10017211 [37] E. Farhi and H. Neven, “Classification with quantum neural networks on near term proces- sors,” arXiv preprint arXiv:.06002, 2018. [38] X. Tang and L. Shu, “Classification of electrocardiogram signals with RS and quantum neu- ral networks,” International Journal of Multimedia Ubiquitous Engineering, vol. 9, no. 2, pp. 363–372, 2014. https://doi.org/10.14257/ijmue.2014.9.2.37 [39] G. Scharmer and M. Carlsson, “A new approach to multi-level non-LTE radiative transfer problems,” Journal of Computational Physics, vol. 59, no. 1, pp. 56–80, 1985. https://doi. org/10.1016/0021-9991(85)90107-X [40] C. Hu, M. Ye, Y. Du, and X. Lu, “Vector projection for face recognition,” Computers Electrical Engineering, vol. 40, no. 8, pp. 51–65, 2014. https://doi.org/10.1016/j. compeleceng.2014.08.010 46 http://www.i-jim.org https://doi.org/10.1002/1097-0355(199023)11:3%3C237::AID-IMHJ2280110306%3E3.0.CO;2-X https://doi.org/10.1002/1097-0355(199023)11:3%3C237::AID-IMHJ2280110306%3E3.0.CO;2-X https://doi.org/10.1111/exsy.12400 https://doi.org/10.1111/exsy.12400 https://doi.org/10.1016/S0020-0255(00)00055-4 https://doi.org/10.1016/S0020-0255(00)00055-4 https://doi.org/10.1006/jcss.2001.1769 https://doi.org/10.1007/978-3-540-45226-3_42 https://doi.org/10.1109/ICNC.2007.70 https://doi.org/10.1109/ICME.2011.6012187 https://doi.org/10.3390/s130506730 https://doi.org/10.3390/s130506730 https://doi.org/10.1504/IJCAT.2018.10017211 https://doi.org/10.14257/ijmue.2014.9.2.37 https://doi.org/10.1016/0021-9991(85)90107-X https://doi.org/10.1016/0021-9991(85)90107-X https://doi.org/10.1016/j.compeleceng.2014.08.010 https://doi.org/10.1016/j.compeleceng.2014.08.010 Paper—Face Patterns Analysis and Recognition System Based on Quantum Neural Network QNN [41] Y. Yan, F. Lee, X. Wu, and Q. Chen, “Face recognition algorithm using extended vector quantization histogram features,” Plos one, vol. 13, no. 1, p. e0190378, 2018. https://doi. org/10.1371/journal.pone.0190378 [42] R. K. Sanodiya, P. Kumar, M. Tiwari, L. Yao, and J. Mathew, “A modified joint geometri- cal and statistical alignment approach for low-resolution face recognition,” in International Conference on Neural Information Processing, 2020, pp. 88–100: Springer. https://doi. org/10.1007/978-3-030-63830-6_8 [43] H.-D. Liu, M. Yang, Y. Gao, Y. Yin, and L. Chen, “Bilinear discriminative dictionary learn- ing for face recognition,” Pattern Recognition, vol. 47, no. 5, pp. 1835–1845, 2014. https:// doi.org/10.1016/j.patcog.2013.11.007 [44] G. Hu, Y. Yang, D. Yi, J. Kittler, W. Christmas, S. Z. Li, and T. Hospedales, “When face recognition meets with deep learning: an evaluation of convolutional neural networks for face recognition,” in Proceedings of the IEEE international conference on computer vision workshops, 2015, pp. 142–150. https://doi.org/10.1109/ICCVW.2015.58 [45] A. Dehghan, E. G. Ortiz, G. Shu, and S. Z. Masood, “Dager: deep age, gender and emotion recognition using convolutional neural network,” arXiv preprint arXiv:.04280, 2017. [46] F. Al-zubidi, N. F. AL-Bakri, R. K. Hasoun, and S. H. Hashim, “Mobile application to detect Covid-19 pandemic by using classification techniques: proposed system,” International Journal of Interactive Mobile Technologies, vol. 15, no. 16, 2021. https://doi.org/10.3991/ ijim.v15i16.24195 [47] J. R. Annam, S. Kalyanapu, S. Ch, J. Somala, and S. B. Raju, “Classification of ECG heart- beat arrhythmia: a review,” Procedia Computer Science, vol. 171, pp. 679–688, 2020. https://doi.org/10.1016/j.procs.2020.04.074 [48] M. Al-dabag, H. S. ALRikabi, and R. Al-Nima, “Anticipating atrial fibrillation signal using efficient algorithm,” International Journal of Online and Biomedical Engineering (iJOE), vol. 17, no. 2, pp. 106–120, 2021. https://doi.org/10.3991/ijoe.v17i02.19183 6 Authors Haider Th. Salim ALRikabi is presently Asst. Prof. and one of the Faculty Col- lege of Engineering, Electrical Engineering Department, Wasit University in Al Kut, Wasit, Iraq. He received his B.Sc. degree in Electrical Engineering in 2006 from the Al Mus-tansiriya University in Baghdad, Iraq. His M.Sc. degree in Electrical Engineer- ing focusing on Communications Systems from California State University/Fullerton/ USA in 2014. His current research interests include Communications systems with the mobile generation, Control systems, intelligent technologies, smart cities, and the Inter- net of Things (IoT). Al Kut City-Hay ALRabee, Wasit, Iraq. E-mail: hdhiyab@uowasit. edu.iq. The number of articles in national databases – 15, The number of articles in international databases – 45. Ibtisam A. Aljazaery is presently Asst. Prof. and on the faculty of Electrical Engi- neering Department, College of Engineering, University of Babylon. Babylon, Iraq. Email: ibtisamalasady@gmail.com. The number of articles in national databases – 10. The number of articles in international databases – 5 Jaafar Sadiq Qateef is a Lecturer in Computer Science Department, College of Education and Pure Sciences, Wasit University, Wasit, Iraq. He completed his master degree at department of information systems from Osmania university, India. Email: Jsadiq@uowasit.edu.iq. iJIM ‒ Vol. 16, No. 08, 2022 47 https://doi.org/10.1371/journal.pone.0190378 https://doi.org/10.1371/journal.pone.0190378 https://doi.org/10.1007/978-3-030-63830-6_8 https://doi.org/10.1007/978-3-030-63830-6_8 https://doi.org/10.1016/j.patcog.2013.11.007 https://doi.org/10.1016/j.patcog.2013.11.007 https://doi.org/10.1109/ICCVW.2015.58 https://doi.org/10.3991/ijim.v15i16.24195 https://doi.org/10.3991/ijim.v15i16.24195 https://doi.org/10.1016/j.procs.2020.04.074 https://doi.org/10.3991/ijoe.v17i02.19183 mailto:hdhiyab@uowasit.edu.iq mailto:hdhiyab@uowasit.edu.iq mailto:ibtisamalasady @gmail.com mailto:Jsadiq@uowasit.edu.iq Paper—Face Patterns Analysis and Recognition System Based on Quantum Neural Network QNN Abdul Hadi M. Alaidi is a Asst. Prof. in the Engineering College, at the Wasit University, Iraq. His area of research focuses on algorithm and image processing. Roa’a M. Al_airaji received the Bachelor degree from Department of Computer, College of Science, University of Babylon, in 2012. Received the Master degree from Department of software, College of Information Technology, University of Babylon, in 2018. Currently work as assistant teacher in College of Science, University of Babylon. Article submitted 2022-02-09. Resubmitted 2022-03-11. Final acceptance 2022-03-11. Final version published as submitted by the authors. 48 http://www.i-jim.org