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

Image compression plays an important role in reducing the size and storage of data while increasing the speed of its transmission through the Internet significantly. Image compression is an important research topic for several decades and recently, with the great successes achieved by deep learning in many areas of image processing, especially image compression, and its use is increasing Gradually in the field of image compression. The deep learning neural network has also achieved great success in the field of processing and compressing various images of different sizes. In this paper, we present a structure for image compression based on the use of a Convolutional AutoEncoder (CAE) for deep learning, inspired by the diversity of human eye

Natural gas and oil are one of the mainstays of the global economy. However, many issues surround the pipelines that transport these resources, including aging infrastructure, environmental impacts, and vulnerability to sabotage operations. Such issues can result in leakages in these pipelines, requiring significant effort to detect and pinpoint their locations. The objective of this project is to develop and implement a method for detecting oil spills caused by leaking oil pipelines using aerial images captured by a drone equipped with a Raspberry Pi 4. Using the message queuing telemetry transport Internet of Things (MQTT IoT) protocol, the acquired images and the global positioning system (GPS) coordinates of the images' acquisition are

Patients infected with the COVID-19 virus develop severe pneumonia, which typically results in death. Radiological data show that the disease involves interstitial lung involvement, lung opacities, bilateral ground-glass opacities, and patchy opacities. This study aimed to improve COVID-19 diagnosis via radiological chest X-ray (CXR) image analysis, making a substantial contribution to the development of a mobile application that efficiently identifies COVID-19, saving medical professionals time and resources. It also allows for timely preventative interventions by using more than 18000 CXR lung images and the MobileNetV2 convolutional neural network (CNN) architecture. The MobileNetV2 deep-learning model performances were evaluated

The main reason for the emergence of a deepfake (deep learning and fake) term is the evolution in artificial intelligence techniques, especially deep learning. Deep learning algorithms, which auto-solve problems when giving large sets of data, are used to swap faces in digital media to create fake media with a realistic appearance. To increase the accuracy of distinguishing a real video from fake one, a new model has been developed based on deep learning and noise residuals. By using Steganalysis Rich Model (SRM) filters, we can gather a low-level noise map that is used as input to a light Convolution neural network (CNN) to classify a real face from fake one. The results of our work show that the training accuracy of the CNN model

Natural settings make it challenging to identify facial expressions since head position, illumination level, and ‎‎occlusion vary. Thus, developing a more generic model without front-facing images alone is quite crucial. This ‎research proposes a facial expression ‎recognition model based on pre-trained deep convolutional neural networks ‎with transfer learning. The model was trained ‎on several cases to classify face expressions into seven ‎classifications efficiently. The proposed system used the EfficientNetB0 model ‎that has one dense dropout layer. The model first rescales and norms the input dataset in the input ‎layer that takes images of a larger resolution to get better results. After entering 7 blocks sequential

Abstract

Hexapod robot is a flexible mechanical robot with six legs. It has the ability to walk over terrain. The hexapod robot look likes the insect so it has the same gaits. These gaits are tripod, wave and ripple gaits. Hexapod robot needs to stay statically stable at all the times during each gait in order not to fall with three or more legs continuously contacts with the ground. The safety static stability walking is called (the stability margin). In this paper, the forward and inverse kinematics are derived for each hexapod’s leg in order to simulate the hexapod robot model walking using MATLAB R2010a for all gaits and the geometry in order to derive the equations of the sub-constraint workspaces for each