China 
Africa 
Explore chapters and articles related to this topic
Object Detection Frameworks and Services in Computer Vision
Published in S. Poonkuntran, Rajesh Kumar Dhanraj, Balamurugan Balusamy, Object Detection with Deep Learning Models, 2023
Sachi Choudhary, Rashmi Sharma, Gargeya Sharma
It can be seen that a perceptron is a linear function, so the trained single neuron will produce a straight line to classify the data. Will this work for complex nonlinear datasets? The answer is that many neurons are needed to optimally fit the training data. A multilayer perceptron has the same structure as a single-layer perceptron but contains two or more hidden layers. Hidden layers are collections of neurons that are not directly accessible by the input data; they act as intermediate processing units between the raw input and the final output. Typically, each neuron in the hidden layer is linked to every other neuron in adjacent layers, forming a denser connection between them and providing more computation to the expected output [4]. Figure 2.3 shows a multilayer perceptron with two hidden layers.
The Foundations of Deep Learning
Published in Hassan Ugail, Deep Learning in Visual Computing, 2022
The perceptron is a classification technique and can be widely utilised in visual computing applications, for example, in computer vision and image recognition. Suppose we have a picture and we want the algorithm to tell us if the picture is that of an apple or not, you can use classification for that. The input could be the pixels of the image, the weights of the perceptron can be derived by using a dataset of apples, and the output will be a binary classification telling us if the input image is an apple or not. However, just like the function, a single neuron is very limited, there are many real-world problems that cannot be solved using binary classifiers. Thus, the perception idea can be extended to artificial neural networks (ANN), where a network of neurons can be connected together.
Alzheimer’s Disease Classification Using Machine Learning Algorithms
Published in J. Dinesh Peter, Steven Lawrence Fernandes, Carlos Eduardo Thomaz, Advances in Computerized Analysis in Clinical and Medical Imaging, 2019
S. Naganandhini, P. Shanmugavadivu, A. Asaithambi, M. Mohammed Mansoor Roomi
A perceptron is an algorithm that classifies given input by separating into two categories with a linear function, and thus called a linear classifier. The input is typically a feature vector x multiplied by weights w and added to a bias b, mathematically expressed as wTx+b. Additionally, a perceptron sometimes passes wTx+b through a nonlinear activation function y and deals with it in the form ywTx+b.
Prediction of wave overtopping discharges at coastal structures using interpretable machine learning
Published in Coastal Engineering Journal, 2023
An ANN has an information-processing structure in the form of a network that connects processors with simple functions on a large scale using the human neural network as a model. A perceptron, which is the basis of deep learning, is a structure built to deliver information according to the threshold by similarly implementing the operation of neurons in brain cells and assigning weight signals to the input values. An NN is divided into an input layer, a hidden layer, and an output layer. Nodes are connected to each other with weights. In an ANN, values in the input layer are delivered to all nodes in the hidden layer using the feedforward method. Moreover, the output values of all nodes in the hidden layer are transmitted to all nodes by combination and activation function calculations. To reduce the difference between the predicted values calculated by this sequence and the experimental values, training is performed using the method of reducing the error by redistributing the weights among all neurons by a backpropagation algorithm.
Architectural space classification considering topological and 3D visual spatial relations using machine learning techniques
Published in Building Research & Information, 2023
Berfin Yıldız, Gülen Çağdaş, Ibrahim Zincir
Artificial neural networks are biologically inspired mathematical systems consisting of many interconnected neurons. In a neural network, each neuron is linked to a mapping that enables it to process multiple inputs and produce a single output. A perceptron refers to a basic network architecture that comprises only an input layer and an output layer. The relationship between the input and output layers in the perceptron is linear. In a perceptron architecture, the input layer transmits its signal directly to the output layer. However, if additional hidden layers are inserted between the input and output layers, the resulting neural network is known as a Multi-Layer Perceptron (MLP). Within the MLP, the weighted signals from the input layer are conveyed to neurons situated within the hidden layers and then processed by the activation function which is nonlinear. MLP network as a deep learning method is also called Feed-Forward Neural Network (FFNN). The FFNN is characterized by fully connected layers, with no inter-neuron connections existing within a given layer.
Stick-Slip Classification Based on Machine Learning Techniques for Building Damage Assessment
Published in Journal of Earthquake Engineering, 2022
Yunsu Na, Sherif El-Tawil, Ahmed Ibrahim, Ahmed Eltawil
The MLP is one of feedforward artificial neural network methods. It consists of more than one perceptron. The perceptron is a simple algorithm to perform binary classification. It is composed of three layers: an input layer, a hidden layer and an output layer. MLP can be tuned to gain better performance depending on the number of layers and nodes in a layer. These two components (the number of layers and nodes in a layer) of the MLP structure are tested in an arbitrary way to seek the optimal combination. From several trials, the MLP with one hidden layer with ten nodes and one output layer with two nodes corresponding to the two classification outputs (sliding and sticking) demonstrate the best performance and are selected in this study.