A three-stage learning algorithm for deep multilayer perceptron (DMLP) with effective weight initialisation based on sparse auto-encoder is proposed in this paper, which aims to overcome difficulties in training deep neural networks with limited training data in high-dimensional feature space. At the first stage, unsupervised learning is adopted using sparse auto-encoder to obtain the initial weights of the feature extraction layers of the DMLP. At the second stage, error back-propagation is used to train the DMLP by fixing the weights obtained at the first stage for its feature extraction layers. At the third stage, all the weights of the DMLP obtained at the second stage are refined by error back-propagation. Network structures an

  With the fast progress of information technology and the computer networks, it becomes very easy to reproduce and share the geospatial data due to its digital styles. Therefore, the usage of geospatial data suffers from various problems such as data authentication, ownership proffering, and illegal copying ,etc. These problems can represent the big challenge to future uses of the geospatial data. This paper introduces a new watermarking scheme to ensure the copyright protection of the digital vector map. The main idea of proposed scheme is based on transforming  the digital map to frequently domain using the Singular Value Decomposition (SVD) in order to determine suitable areas to insert the watermark data.

Purpose:  To compare the central corneal thickness (CCT),minimum corneal thickness (MCT) and corneal power measured using theScheimpflug-Placido device and optical coherence tomography (OCT) in healthy eyes. Study Design:  Descriptive observational. Place and Duration of Study:  Al-Kindy college of medicine/university of Baghdad, from June 2021 to April 2022. Methods:  A total of 200 eyes of 200 individuals were enrolled in this study. CCT and MCT measurements were carried out using spectral-domain optical coherence tomography (Optovue) and a Scheimpflug-Placido topographer (Sirius).The agreement between the two approaches was assessed using Bland-Altman analysis in this study. Results:  Mean age was 28.54 ± 6.6 years, me

BACKGROUND: Keratoconus is a progressive non inflammatory bilateral (usually asymmetric) ectatic corneal disease characterized by paraxial stromal thinning ,weakening that lead to corneal surface distortion ,vision loss primarily from irregular astigmatism and myopia and secondly from corneal scar. OBJECTIVE: To evaluate visual and refractive outcomes after intracorneal continuous ring (ICCR) implantation combined with intrapocket corneal collagen cross linking in patient with keratoconus. Setting: Eye Specialty Private Hospital, Baghdad, Iraq. METHODS: This study assessed the results of implantation of Myoring ICCR combined with CXL in 40 eyes with KC. Outcome measures include UDVA,CDVA(spectacle correction),refraction, complications and s

Diagnosing heart disease has become a very important topic for researchers specializing in artificial intelligence, because intelligence is involved in most diseases, especially after the Corona pandemic, which forced the world to turn to intelligence. Therefore, the basic idea in this research was to shed light on the diagnosis of heart diseases by relying on deep learning of a pre-trained model (Efficient b3) under the premise of using the electrical signals of the electrocardiogram and resample the signal in order to introduce it to the neural network with only trimming processing operations because it is an electrical signal whose parameters cannot be changed. The data set (China Physiological Signal Challenge -cspsc2018) was ad

In this article, Convolution Neural Network (CNN) is used to detect damage and no damage images form satellite imagery using different classifiers. These classifiers are well-known models that are used with CNN to detect and classify images using a specific dataset. The dataset used belongs to the Huston hurricane that caused several damages in the nearby areas. In addition, a transfer learning property is used to store the knowledge (weights) and reuse it in the next task. Moreover, each applied classifier is used to detect the images from the dataset after it is split into training, testing and validation. Keras library is used to apply the CNN algorithm with each selected classifier to detect the images. Furthermore, the performa

— To identify the effect of deep learning strategy on mathematics achievement and practical intelligence among secondary school students during the 2022/2023 academic year. In the research, the experimental research method with two groups (experimental and control) with a post-test were adopted. The research community is represented by the female students of the fifth scientific grade from the first Karkh Education Directorate. (61) female students were intentionally chosen, and they were divided into two groups: an experimental group (30) students who were taught according to the proposed strategy, and a control group (31) students who were taught according to the usual method. For the purpose of collecting data for the experimen

    The Internet of Things (IoT) technology and smart systems are playing a major role in the advanced developments in the world that take place nowadays, especially in multiple privilege systems. There are many smart systems used in daily human life to serve them and facilitate their tasks, such as alarm systems that work to prevent unwanted events or face detection and recognition systems. The main idea of this work is to capture live video using a connected Pi camera, save it, and unlock the electric strike door in several ways; either automatically by displaying a live video connected via USB webcam using a deep learning algorithm of facial recognition and OpenCV or by RFID technology, as well as by detecting abnormal entrance wit

    In data mining and machine learning methods, it is traditionally assumed that training data, test data, and the data that will be processed in the future, should have the same feature space distribution. This is a condition that will not happen in the real world. In order to overcome this challenge, domain adaptation-based methods are used. One of the existing challenges in domain adaptation-based methods is to select the most efficient features so that they can also show the most efficiency in the destination database. In this paper, a new feature selection method based on deep reinforcement learning is proposed. In the proposed method, in order to select the best and most appropriate features, the essential policies