Keeping less valid data to obtain necessary information has become a new requirement in the signal-processing field. The paper employs adaptive dictionary for sparse representation, introduces a characteristic-weighti...
详细信息
ISBN:
(纸本)9781538660058
Keeping less valid data to obtain necessary information has become a new requirement in the signal-processing field. The paper employs adaptive dictionary for sparse representation, introduces a characteristic-weighting coefficient to offer detailed image information, and meanwhile performs Schmidt orthogonalization with the combination of Gaussian random measurement matrix to minimize the correlation of vectors in matrix. It raises the figure structural group sparse representation (FSGSR) algorithm based on matrix orthogonalization. Experiments indicate that this improved image reconstruction algorithm has enhanced the reconstructed image quality compared with typical algorithms during same time length.
Acute Lymphoblastic Leukemia (ALL) is the most prevalent acute leukemia in adults after Acute Myeloid Leukemia, with a diffusion of over 6500 persons per year just in the United States. In this research, we propose a ...
详细信息
ISBN:
(纸本)9781538611227
Acute Lymphoblastic Leukemia (ALL) is the most prevalent acute leukemia in adults after Acute Myeloid Leukemia, with a diffusion of over 6500 persons per year just in the United States. In this research, we propose a smart assistant determination method for ALL diagnosis using microscopic images. In this regard, K-means is employed to extract cell images after that wavelet transform is hired on cell images then statistical moments of the transformed image are computed to extract features. Afterward, a Chain Tabu search algorithm is proposed for feature selection of normal and abnormal cells to enable classifiers classifying ALL efficiently. Finally, Multi-Layer Perceptron (MLP) is used for classification. The proposed method is evaluated on ALL-IDB2. The proposed method achieved the accuracy of 98.88% and outperforms existed ALL diagnosis methods.
This paper presents the use of hyperspectral imageprocessing as an alternative to traditional mineral exploration techniques for identification of regions in the study area rich in zinc mineral. Zincian Dolomite is a...
详细信息
ISBN:
(数字)9781728112619
ISBN:
(纸本)9781728112626
This paper presents the use of hyperspectral imageprocessing as an alternative to traditional mineral exploration techniques for identification of regions in the study area rich in zinc mineral. Zincian Dolomite is a known ore of zinc found in the study area. The Hyperion sensor dataset of a study area in Rajasthan, India is obtained from the USGS (United States Geological Survey) Earth explorer and pre-processed using ENvi (Environment for visualizing images) software to remove noise. The pre-processed data is later exported to a dedicated python program which uses machine learning algorithms to identify the regions rich in Zincian Dolomite. The proposed technique form mineral exploration is faster and cheaper than the traditional techniques and is also found to be very accurate in identification of the minerals.
In this paper, an image sharpening method using integral image representation and Laplacian operator is presented. First, a parallel algorithm is proposed to compute the integral image of the original image. Then, the...
详细信息
ISBN:
(纸本)9781538663011
In this paper, an image sharpening method using integral image representation and Laplacian operator is presented. First, a parallel algorithm is proposed to compute the integral image of the original image. Then, the integral image is used to compute the Laplacian image by subtracting the center pixel from its surround average in a rectangular window. This method can achieve a constant number of operations per rectangle. Next, the sharpened image is obtained by adding the Laplacian image to the original image. Finally, one numerical example is demonstrated to show the effectiveness of the proposed image sharpening approach.
The quality of image collected under severe weather such as fog and haze is badly damaged due to the atmospheric scattering. In order to solve the problem that image dehazing algorithms have poor adaptability, which w...
详细信息
ISBN:
(纸本)9781538671740
The quality of image collected under severe weather such as fog and haze is badly damaged due to the atmospheric scattering. In order to solve the problem that image dehazing algorithms have poor adaptability, which will occur contrast distortion after restoration or cannot eliminating the influence of dense haze, a self-adaption single image dehaze method based on clarity evaluation is proposed to effectively recover the visual effects of the scene. The innovation points of this paper lied in that, first, the haze image is disposed separately according to average value, standard deviation, average gradient, information entropy and other clarity judgment features of the input image;then, the method of self-adaption image quality evaluation and coding decision is introduced to the dehaze results, to output the best effects obtained through comparison of many methods;finally, the clarity judgment is carried out again to output the final results. Experimental results demonstrate that the proposed method can achieve a better dehazing effect, and its chromaticity, luminance and contrast are improved to a certain extent. The universality of dehaze method is further improved.
In this work, we propose an ultra-fast unsupervised saliency detection algorithm based on QR matrix factorization. The algorithm works by dividing an image into blocks and computing the QR factors in each local neighb...
详细信息
ISBN:
(数字)9781728143002
ISBN:
(纸本)9781728143019
In this work, we propose an ultra-fast unsupervised saliency detection algorithm based on QR matrix factorization. The algorithm works by dividing an image into blocks and computing the QR factors in each local neighborhood where salient patches correspond to columns in R with largest L 0 norm. Experimental results on the MSRA10K salient object database show that the proposed algorithm is competitive with state-of-the-art real-time saliency detection algorithms in terms of AUC and execution time. Additionally, we study the various parameters that control the performance of the proposed algorithm such as threshold values and processing-block sizes and their effect on the overall performance.
The paper constitutes the implementation of Convolutional neural network for the emotion detection and thereby playing a song accordingly. Segregating the songs and playing them in accordance to one's mood could f...
详细信息
ISBN:
(纸本)9781728116006
The paper constitutes the implementation of Convolutional neural network for the emotion detection and thereby playing a song accordingly. Segregating the songs and playing them in accordance to one's mood could facilitate the music lover. Although there exist a lot of algorithms designed for it, the computation is not as expected. This paper eradicates such an issue by using CNN. In order to obtain minimal processing, multilayer perceptron are implemented by CNNs. In comparison to various algorithms for image classification, CNNs observed to have little-processing. This implies that the filters used in CNNs are advantageous when compared to traditional algorithm. The visualization of features directly can be less informative. Hence, we use the training procedure of back-propagation to activate the filters for better visualization. The multiple actions such as capturing, detecting the emotion and classifying the same can all be confined as one step through the use of CNN. The slow performances of the real-time approaches could be enhanced by regularizing the methods and by visualizing the hidden features. Hence the proposed approach could enhance the accuracy and the computation speed.
Currently, Intelligent driver assistance systems provide more satisfactory solutions to the field of road safety. These are based on the use of powerful image or signal processingalgorithms in terms of speed, accurac...
详细信息
ISBN:
(数字)9781728140582
ISBN:
(纸本)9781728140599
Currently, Intelligent driver assistance systems provide more satisfactory solutions to the field of road safety. These are based on the use of powerful image or signal processingalgorithms in terms of speed, accuracy of calculation and decision making. This document describes an intelligent roadway classification system in real-time in 4 classes: asphalt, gravel, snow and stone road. This system relies on the use and combination of two algorithms; the Mel Frequency Cepstrum Coefficient algorithm that extracts the characteristics of each type of road from the sound due to friction of the vehicle wheels with the road, and Multilayer Perceptron algorithm, which receives its characteristics to its inputs and classifies them. This document also describes the stages of the process of learning and testing. The robustness of this system is shown by the obtained satisfactory experimental results that combine between the respective advantages of the MFCC descriptor and MLP classifier.
The proceedings contain 47 papers. The special focus in this conference is on Smart Innovations in Communications and Computational Sciences. The topics include: Computer vision-Based Fruit Disease Detection and Class...
ISBN:
(纸本)9789811324130
The proceedings contain 47 papers. The special focus in this conference is on Smart Innovations in Communications and Computational Sciences. The topics include: Computer vision-Based Fruit Disease Detection and Classification;Automated Brain Tumor Detection Using Discriminative Clustering Based MRI Segmentation;cancer Prediction Based on Fuzzy Inference System;energy Consumption Data Analysis and Operation Evaluation of Green Buildings;the Method of Random Generation of Electronic Patrol Path Based on Artificial Intelligence;Channel Power Estimation for DVBRCS to DVBS2 Onboard DSP Payload;a Slotted Microstrip Patch Antenna for 5G Mobile Phone Applications;reliability Study of Sensor Node Monitoring Unattended Environment;RSOM-Based Clustering and Routing in WSNs;real-Time Meta Learning Approach for Mobile Healthcare;Evaluation of Received Signal Strength Indicator (RSSI) for Relay-Based Communication in WBAN;side-Channel Attacks on Cryptographic Devices and Their Countermeasures—A Review;implementation and Analysis of Different Path Loss Models for Cooperative Communication in a Wireless Sensor Network;Parallel Approach for Sub-graph Isomorphism on Multicore System Using OpenMP;optimized Solution for Employee Transportation Problem Using Linear Programming;spectrum Prediction Using Time Delay Neural Network in Cognitive Radio Network;Reliability Factor Based AODV Protocol: Prevention of Black Hole Attack in MANET;a Survey of Lightweight Cryptographic algorithms for IoT-Based Applications;fruit Disease Detection Using Rule-Based Classification;Fault Tolerance Through Energy Balanced Cluster Formation (EBCF) in WSN;firearm Detection from Surveillance Cameras Using imageprocessing and Machine Learning Techniques;Mining Social Networks: Tollywood Reviews for Analyzing UPC by Using Big Data Framework;cloud-Based E-Learning: Using Cloud Computing Platform for an Effective E-Learning.
Today images and videos are everywhere. In fact, the sheer quantity of images on social media and networking sites is unfathomable. Every device is now fitted with a camera. This opens up huge possibilities. Object Re...
详细信息
Today images and videos are everywhere. In fact, the sheer quantity of images on social media and networking sites is unfathomable. Every device is now fitted with a camera. This opens up huge possibilities. Object Recognition is a process of detecting an object and identifying it using various imagealgorithms. The main purpose of this paper is to recognize objects in real time and allot the objects to the classes that are previously defined. The algorithms that we utilized are more computationally efficient. Previously, object detection was done using RFID and IR technologies which required dedicated hardware. But with the advent of imageprocessing and neural networks, we require almost no new hardware. Almost everything has camera these days from pens to mobile phones. This has given rise to a new field called computer vision i.e. using pictures and videos to detect, segregate and track objects or events so that we can “understand” a real world scenario.
暂无评论