Tumor detection in colonoscopic images using hybrid methods for on-line neural network training
نویسندگان
چکیده
In this paper the effectiveness of a new Hybrid Evolutionary Algorithm in on-line Neural Network training for tumor detection is investigated. To this end, a Lamarck-inspired combination of Evolutionary Algorithms and Stochastic Gradient Descent is proposed. The Evolutionary Algorithm works on the termination point of the Stochastic Gradient Descent. Thus, the method consists in a Stochastic Gradient Descent-based on-line training stage and an Evolutionary Algorithm-based retraining stage. On-line training is considered eminently suitable for large (or even redundant) training sets and/or networks; it also helps escaping local minima and provides a more natural approach for learning nonstationary tasks. Furthermore, the notion of retraining aids the hybrid method to exhibit reliable and stable performance, and increases the generalization capability of the trained neural network. Experimental results suggest that the proposed hybrid strategy is capable to train on-line, efficiently and effectively. Here, an artificial neural network architecture has been successfully used for detecting abnormalities in colonoscopic video images.
منابع مشابه
Neural network-based colonoscopic diagnosis using on-line learning and differential evolution
In this paper, on-line training of neural networks is investigated in the context of computerassisted colonoscopic diagnosis. A memory-based adaptation of the learning rate for the on-line Backpropagation is proposed and used to seed an on-line evolution process that applies a Differential Evolution Strategy to (re-)adapt the neural network to modified environmental conditions. Our approach loo...
متن کاملNeural Network Performance Analysis for Real Time Hand Gesture Tracking Based on Hu Moment and Hybrid Features
This paper presents a comparison study between the multilayer perceptron (MLP) and radial basis function (RBF) neural networks with supervised learning and back propagation algorithm to track hand gestures. Both networks have two output classes which are hand and face. Skin is detected by a regional based algorithm in the image, and then networks are applied on video sequences frame by frame in...
متن کاملFace Detection with methods based on color by using Artificial Neural Network
The face Detection methodsis used in order to provide security. The mentioned methods problems are that it cannot be categorized because of the great differences and varieties in the face of individuals. In this paper, face Detection methods has been presented for overcoming upon these problems based on skin color datum. The researcher gathered a face database of 30 individuals consisting of ov...
متن کاملNon-melanoma skin cancer diagnosis with a convolutional neural network
Background: The most common types of non-melanoma skin cancer are basal cell carcinoma (BCC), and squamous cell carcinoma (SCC). AKIEC -Actinic keratoses (Solar keratoses) and intraepithelial carcinoma (Bowen’s disease)- are common non-invasive precursors of SCC, which may progress to invasive SCC, if left untreated. Due to the importance of early detection in cancer treatment, this study aimed...
متن کاملA Two-Dimensional Convolutional Neural Network for Brain Tumor Detection From MRI
Aims: Cancerous brain tumors are among the most dangerous diseases that lower the quality of life of people for many years. Their detection in the early stages paves the way for the proper treatment. The present study aimed to present a two-dimensional Convolutional Neural Network (CNN) for detecting brain tumors under Magnetic Resonance Imaging (MRI) using the deep learning method. Methods & ...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
عنوان ژورنال:
دوره شماره
صفحات -
تاریخ انتشار 2001