Modified Dendrite Morphological Neural Network Applied to 3D Object Recognition

Abstract

In this paper a modified dendrite morphological neural network (DMNN) is applied for recognition and classification of 3D objects. For feature extraction, the first two Hu's moment invariants are calculated based on 2D binary images, as well as the mean and the standard deviation obtained on 2D grayscale images. These four features were fed into a DMNN for classification of 3D objects. For testing, COIL-20 image database and a generated dataset were used. A comparative analysis of the proposed method with MLP and SVM is presented and the results reveal the advantages of the modified DMNN. An important characteristic of the proposed recognition method is that because of the simplicity of calculation of the extracted features and the DMNN, this method can be used in real applications.

A Dendritic Neuron Model with Adaptive Synapses Trained by Differential Evolution Algorithm

“…In this section, we compared DE-DMAS with the most popular model BPNN. To make the comparison relatively fair, the number of adjusted weights and thresholds (D BPNN and D DE− OMAS shown in equations (10) and (11), respectively) in both models should be arranged nearly the same because these numbers generally determine the size of the model and the computational complexity although the two models have different architectures:…”

“…In recent years, several dendritic computing models considering the functions of dendrites in a neuron have been proposed in the literature. A dendritic morphological neural network (DMNN) which is based on the traditional morphological neural networks [8,9] is proposed for solving classification problems [10] and 3D object recognition tasks [11]. A nonlinear dendritic neuron model equipped with binary synapses [11] is demonstrated to be capable of learning temporal features of spike input patterns.…”

“…A dendritic morphological neural network (DMNN) which is based on the traditional morphological neural networks [8,9] is proposed for solving classification problems [10] and 3D object recognition tasks [11]. A nonlinear dendritic neuron model equipped with binary synapses [11] is demonstrated to be capable of learning temporal features of spike input patterns. Most recently, a dendritic neuron model (DNM) with nonlinear synapses has been proposed [12][13][14].…”

Training Algorithm for Dendrite Morphological Neural Network Using K-Medoids

“…Neural Networks (DMNN) as a new training method based on a Divide and Conquer strategy in [24]. The application was used to recognize 3D objects using Kinect [25]. Pessoa and Maragos in [26] combine classical perceptrons with morphological/rank neurons called Morphological/rank/linear neural networks (MRL-NNs).…”

Efficient Lane Detection Based on Artificial Neural Networks

“…Interest in Morphological Neural Networks (MNNs) has increased in the last five years (Sossa, 2013), (Sossa, 2014), (Arce, 2016) and . This kind of ANNs make use of min and max operations to segment the input space into hyperboxes.…”