Refine Search

New Search

Results: 38

(searched for: doi:10.13176/11.191)
Save to Scifeed
Page of 1
Articles per Page
by
Show export options
  Select all
Published: 21 November 2021
by MDPI
Remote Sensing, Volume 13; https://doi.org/10.3390/rs13224712

Abstract:
Image classification has always been a hot research direction in the world, and the emergence of deep learning has promoted the development of this field. Convolutional neural networks (CNNs) have gradually become the mainstream algorithm for image classification since 2012, and the CNN architecture applied to other visual recognition tasks (such as object detection, object localization, and semantic segmentation) is generally derived from the network architecture in image classification. In the wake of these successes, CNN-based methods have emerged in remote sensing image scene classification and achieved advanced classification accuracy. In this review, which focuses on the application of CNNs to image classification tasks, we cover their development, from their predecessors up to recent state-of-the-art (SOAT) network architectures. Along the way, we analyze (1) the basic structure of artificial neural networks (ANNs) and the basic network layers of CNNs, (2) the classic predecessor network models, (3) the recent SOAT network algorithms, (4) comprehensive comparison of various image classification methods mentioned in this article. Finally, we have also summarized the main analysis and discussion in this article, as well as introduce some of the current trends.
Joseph Agyapong Mensah, , Felix O. Mettle,
Published: 20 September 2021
Journal of Applied Mathematics, Volume 2021, pp 1-12; https://doi.org/10.1155/2021/7060270

Abstract:
Many architectures of face recognition modules have been developed to tackle the challenges posed by varying environmental constraints such as illumination, occlusions, pose, and expressions. These recognition systems have mainly focused on a single constraint at a time and have achieved remarkable successes. However, the presence of multiple constraints may deteriorate the performance of these face recognition systems. In this study, we assessed the performance of Principal Component Analysis and Singular Value Decomposition using Discrete Wavelet Transform (DWT-PCA/SVD) for preprocessing face recognition algorithm on multiple constraints (partially occluded face images acquired with varying expressions). Numerical evaluation of the study algorithm gave reasonably average recognition rates of 77.31% and 76.85% for left and right reconstructed face images with varying expressions, respectively. A statistically significant difference was established between the average recognition distance of the left and right reconstructed face images acquired with varying expressions using pairwise comparison test. The post hoc analysis using the Bonferroni simultaneous confidence interval revealed that the significant difference established through the pairwise comparison test was mainly due to the sad expressions. Although the performance of the DWT-PCA/SVD algorithm declined as compared to its performance on single constraints, the algorithm attained appreciable performance level under multiple constraints. The DWT-PCA/SVD recognition algorithm performs reasonably well for recognition when partial occlusion with varying expressions is the underlying constraint.
, Joseph Agyapong Mensah, Francis Ayiah-Mensah, Felix O. Mettle
Published: 15 June 2021
Advances in Multimedia, Volume 2021, pp 1-11; https://doi.org/10.1155/2021/4981394

Abstract:
The drift towards face-based recognition systems can be attributed to recent advances in supportive technology and emerging areas of application including voting systems, access control, human-computer interactions, entertainments, and crime control. Despite the obvious advantages of such systems being less intrusive and requiring minimal cooperation of subjects, the performances of their underlying recognition algorithms are challenged by the quality of face images, usually acquired from uncontrolled environments with poor illuminations, varying head poses, ageing, facial expressions, and occlusions. Although several researchers have leveraged on the property of bilateral symmetry to reconstruct half-occluded face images, their approach becomes deficient in the presence of random occlusions. In this paper, we harnessed the benefits of the multiple imputation by the chained equation technique and image denoising using Discrete Wavelet Transforms (DWTs) to reconstruct degraded face images with random missing pixels. Numerical evaluation of the study algorithm gave a perfect (100%) average recognition rate each for recognition of occluded and augmented face images. The study also revealed that the average recognition rate for the augmented face images (75.5811) was significantly lower than the average recognition rate (430.7153) of the occluded face images. MICE augmentation is recommended as a suitable data enhancement mechanism for imputing missing data/pixel of occluded face images.
Samwel Opiyo, Cedric Okinda, , Emmy Mwangi, Nelson Makange
Computers and Electronics in Agriculture, Volume 185; https://doi.org/10.1016/j.compag.2021.106153

The publisher has not yet granted permission to display this abstract.
Sunanda Das, Sourav De,
Research Anthology on Advancements in Quantum Technology pp 164-196; https://doi.org/10.4018/978-1-7998-8593-1.ch007

Abstract:
In this chapter, a quantum-induced modified-genetic-algorithm-based FCM clustering approach is proposed for true color image segmentation. This approach brings down the early convergence problem of FCM to local minima point, increases efficacy of conventional genetic algorithm, and decreases the computational cost and execution time. Effectiveness of genetic algorithm is tumid by modifying some features in population initialization and crossover section. To speed up the execution time as well as make it cost effective and also to get more optimized class levels some quantum computing phenomena like qubit, superposition, entanglement, quantum rotation gate are induced to modified genetic algorithm. Class levels which are yield now fed to FCM as initial input class levels; thus, the ultimate segmented results are formed. Efficiency of proposed method are compared with classical modified-genetic-algorithm-based FCM and conventional FCM based on some standard statistical measures.
Zhiyong Liu, Chuan Yang, Jun Huang, Shaopeng Liu, Yumin Zhuo,
Published: 20 August 2020
Future Generation Computer Systems, Volume 114, pp 358-367; https://doi.org/10.1016/j.future.2020.08.015

The publisher has not yet granted permission to display this abstract.
A A Egorchev
IOP Conference Series: Materials Science and Engineering, Volume 873; https://doi.org/10.1088/1757-899x/873/1/012007

Abstract:
In this article, the algorithm of preliminary filtering of images of a video stream based on assessment and the subsequent correction of such qualitative characteristics of images as contrast, noise level, sharpness and existence of illumination is considered.
Gaurang Bansal, , Pratik Narang, Subham Kumar, Sundaresan Raman
IET Image Processing, Volume 14, pp 1240-1247; https://doi.org/10.1049/iet-ipr.2019.1164

Abstract:
With the increasing incidence rate of lung cancer patients, early diagnosis could help in reducing the mortality rate. However, accurate recognition of cancerous lesions is immensely challenging owing to factors such as low contrast variation, heterogeneity and visual similarity between benign and malignant nodules. Deep learning techniques have been very effective in performing natural image segmentation with robustness to previously unseen situations, reasonable scale invariance and the ability to detect even minute differences. However, they usually fail to learn domain-specific features due to the limited amount of available data and domain agnostic nature of these techniques. This work presents an ensemble framework Deep3DSCan for lung cancer segmentation and classification. The deep 3D segmentation network generates the 3D volume of interest from computed tomography scans of patients. The deep features and handcrafted descriptors are extracted using a fine-tuned residual network and morphological techniques, respectively. Finally, the fused features are used for cancer classification. The experiments were conducted on the publicly available LUNA16 dataset. For the segmentation, the authors achieved an accuracy of 0.927, significant improvement over the template matching technique, which had achieved an accuracy of 0.927. For the detection, previous state-of-the-art is 0.866, while ours is 0.883.
, Katrin S. Lohan
Published: 21 January 2020
Frontiers in Robotics and AI, Volume 6; https://doi.org/10.3389/frobt.2019.00154

Abstract:
Most collaborative tasks require interaction with everyday objects (e.g., utensils while cooking). Thus, robots must perceive everyday objects in an effective and efficient way. This highlights the necessity of understanding environmental factors and their impact on visual perception, such as illumination changes throughout the day on robotic systems in the real world. In object recognition, two of these factors are changes due to illumination of the scene and differences in the sensors capturing it. In this paper, we will present data augmentations for object recognition that enhance a deep learning architecture. We will show how simple linear and non-linear illumination models and feature concatenation can be used to improve deep learning-based approaches. The aim of this work is to allow for more realistic Human-Robot Interaction scenarios with a small amount of training data in combination with incremental interactive object learning. This will benefit the interaction with the robot to maximize object learning for long-term and location-independent learning in unshaped environments. With our model-based analysis, we showed that changes in illumination affect recognition approaches that use Deep Convolutional Neural Network to encode features for object recognition. Using data augmentation, we were able to show that such a system can be modified toward a more robust recognition without retraining the network. Additionally, we have shown that using simple brightness change models can help to improve the recognition across all training set sizes.
Published: 12 February 2019
Informatics in Medicine Unlocked, Volume 14, pp 23-33; https://doi.org/10.1016/j.imu.2019.02.001

Abstract:
Globally, cervical cancer ranks as the fourth most prevalent cancer affecting women. However, it can be successfully treated if detected at an early stage. The Pap smear is a good tool for initial screening of cervical cancer, but there is the possibility of error due to human mistake. Moreover, the process is tedious and time-consuming. The objective of this study was to mitigate the risk of mistake by automating the process of cervical cancer classification from Pap smear images. In this research, contrast local adaptive histogram equalization was used for image enhancement. Cell segmentation was achieved through a Trainable Weka Segmentation classifier, and a sequential elimination approach was used for debris rejection. Feature selection was achieved using simulated annealing integrated with a wrapper filter, while classification was achieved using a fuzzy c-means algorithm. The evaluation of the classifier was carried out on three different datasets (single cell images, multiple cell images and Pap smear slide images from a pathology unit). An overall classification accuracy, sensitivity and specificity of ‘98.88%, 99.28% and 97.47%’, ‘97.64%, 98.08% and 97.16%’ and ‘96.80%, 98.40% and 95.20%’ were obtained for each dataset respectively. The higher accuracy and sensitivity of the classifier was attributed to the robustness of the feature selection method that was utilized to select cell features that would improve the classification performance, and the number of clusters used during defuzzification and classification. The evaluation and testing conducted confirmed the rationale of the approach taken, which is based on the premise that the selection of salient features embeds sufficient discriminatory information that leads to an increase in the accuracy of cervical cancer classification. Results show that the method outperforms many of the existing algorithms in terms of the false negative rate (0.72%), false positive rate (2.53%), and classification error (1.12%), when applied to the DTU/Herlev benchmark Pap smear dataset. The approach articulated in this paper is applicable to many Pap smear analysis systems, but is particularly pertinent to low-cost systems that should be of significant benefit to developing economies.
, Omar Ramadan
2018 9th International Conference on Information and Communication Systems (ICICS) pp 111-116; https://doi.org/10.1109/iacs.2018.8355451

Abstract:
In this paper, a new re-ranking approach based on the multimedia contents and some user specific actions is presented for improving the performance of multimedia search engines. The proposed approach has the ability of working with all multimedia types: video, image, and audio. In addition, a group of descriptors with weight assigned dynamically have been used to describe the multimedia files accurately. Furthermore, the weight of each descriptor, which affects the file rank, is assigned based on the percentage of differences that found by the descriptor. Several experiments have been conducted and it has been observed that the proposed re-ranking approach shows the most relevant files to the top of the query results, and increases the percentage of the retrieved relevant files.
Advances in Computer and Electrical Engineering pp 55-94; https://doi.org/10.4018/978-1-5225-5219-2.ch003

Abstract:
In this chapter, a quantum-induced modified-genetic-algorithm-based FCM clustering approach is proposed for true color image segmentation. This approach brings down the early convergence problem of FCM to local minima point, increases efficacy of conventional genetic algorithm, and decreases the computational cost and execution time. Effectiveness of genetic algorithm is tumid by modifying some features in population initialization and crossover section. To speed up the execution time as well as make it cost effective and also to get more optimized class levels some quantum computing phenomena like qubit, superposition, entanglement, quantum rotation gate are induced to modified genetic algorithm. Class levels which are yield now fed to FCM as initial input class levels; thus, the ultimate segmented results are formed. Efficiency of proposed method are compared with classical modified-genetic-algorithm-based FCM and conventional FCM based on some standard statistical measures.
Jia-Mei Chen, Yan Li, Jun Xu, Lei Gong, Lin-Wei Wang, Wen-Lou Liu, Juan Liu
Published: 28 March 2017
Tumor Biology, Volume 39; https://doi.org/10.1177/1010428317694550

Abstract:
With the advance of digital pathology, image analysis has begun to show its advantages in information analysis of hematoxylin and eosin histopathology images. Generally, histological features in hematoxylin and eosin images are measured to evaluate tumor grade and prognosis for breast cancer. This review summarized recent works in image analysis of hematoxylin and eosin histopathology images for breast cancer prognosis. First, prognostic factors for breast cancer based on hematoxylin and eosin histopathology images were summarized. Then, usual procedures of image analysis for breast cancer prognosis were systematically reviewed, including image acquisition, image preprocessing, image detection and segmentation, and feature extraction. Finally, the prognostic value of image features and image feature–based prognostic models was evaluated. Moreover, we discussed the issues of current analysis, and some directions for future research.
Sandip Dey, Siddhartha Bhattacharyya,
Published: 1 January 2017
Nature-Inspired Computing pp 893-931; https://doi.org/10.4018/978-1-5225-0788-8.ch034

Abstract:
In this chapter, an exhaustive survey of quantum behaved techniques on swarm intelligent is presented. The techniques have been categorized into different classes, and in conclusion, a comparison is made according to the benefits of the approaches taken for review. The above-mentioned techniques are classified based on the information they exploit, for instance, neural network related, meta-heuristic and evolutionary algorithm related, and other distinguished approaches are considered. Neural Network-Based Approaches exhibit a few brain-like activities, which are programmatically complicated, for instance, learning, optimization, etc. Meta-Heuristic Approaches update solution generation-wise for optimization, and the approaches differ based on the problem definition.
Jose Luis Seixas, Rafael G. Mantovani
2016 International Conference on Computational Science and Computational Intelligence (CSCI) pp 677-681; https://doi.org/10.1109/csci.2016.0133

Abstract:
Misleading diagnosis of skin diseases can result in complications during the healing process. Skin images provide important information for the medical staff for information storage and exchange, to trying to prevent this misdiagnosis from happening. For such, a good segmentation process is needed. The segmentation of these images is already being used and has been an effective tool for skin diseases recognition. This paper presents a method for targeting seeds for region growing algorithms, as several of region growing algorithms have good clustering results, but are sensitive to seed. Machine learning were use to create the seed for segmentation of medical images of skin ulcers in the lower limbs. For machine learning, decision tree algorithms were used, which bring a more intuitive approach. The results were compared with gold standard obtained with the help of experts, the results were good and opened paths that can be followed for further work since, even though good results, they can still be improved.
Xiaohui Wang, Jingyan Qin, Yujiao Gao
International Journal of Pattern Recognition and Artificial Intelligence, Volume 30; https://doi.org/10.1142/s0218001416540057

The publisher has not yet granted permission to display this abstract.
Handbook of Research on Machine Learning Innovations and Trends pp 321-348; https://doi.org/10.4018/978-1-4666-9474-3.ch011

Abstract:
A self-supervised image segmentation method by a non-dominated sorting genetic algorithm-II (NSGA-II) based optimized MUSIG (OptiMUSIG) activation function with a multilayer self-organizing neural network (MLSONN) architecture is proposed to segment multilevel gray scale images. In the same way, another NSGA-II based parallel version of the OptiMUSIG (ParaOptiMUSIG) activation function with a parallel self-organizing neural network (PSONN) architecture is purported to segment the color images in this article. These methods are intended to overcome the drawback of their single objective based counterparts. Three standard objective functions are employed as the multiple objective criteria of the NSGA-II algorithm to measure the quality of the segmented images.
Handbook of Research on Machine Learning Innovations and Trends pp 349-377; https://doi.org/10.4018/978-1-4666-9474-3.ch012

Abstract:
In this article, a genetic algorithm inspired by quantum computing is presented. The novel algorithm referred to as quantum inspired genetic algorithm (QIGA) is applied to determine optimal threshold of two gray level images. Different random chaotic map models exhibit the inherent interference operation in collaboration with qubit and superposition of states. The random interference is followed by three different quantum operators viz., quantum crossover, quantum mutation and quantum shifting produce population diversity. Finally, the intermediate states pass through the quantum measurement for optimization of image thresholding. In the proposed algorithm three evaluation metrics such as Brinks's, Kapur's and Pun's algorithms have been applied to two gray level images viz., Lena and Barbara. These algorithms have been applied in conventional GA and Han et al.'s QEA. A comparative study has been made between the proposed QIGA, Han et al.'s algorithm and conventional GA that indicates encouraging avenues of the proposed QIGA.
Siddhartha Bhattacharyya, Pankaj Pal, Sandip Bhowmick
2015 Fifth International Conference on Communication Systems and Network Technologies pp 1129-1135; https://doi.org/10.1109/csnt.2015.55

Abstract:
In this article, a quantum version of the parallel self organizing neural network (QPSONN) architecture for extraction of pure color objects from a noisy perspective is proposed. The QPSONN architecture operates in a phased manner to process input noisy pure color images. After the segregation of the pure color inputs into pure color components in the initial phase, these components are subsequently forwarded for processing to three component quantum multilayer self organizing neural network (QMLSONN) architectures composed of three processing layers viz., input, hidden and output layers characterized by qubits based neurons. The interconnection weights are represented by single qub it rotation gates. Quantum measurements at the component output layers destroy the quantum states of the processed information facilitating adjustment of network interconnection weights by a quantum back propagation algorithm using linear indices of fuzziness. Finally, a fusion of the stable component outputs are brought about in a sink layer to produce extracted outputs. Results of application of the QPSONN are demonstrated on a synthetic and a real life spanner image with various degrees of Gaussian noise. A comparison with the classical PSONN architecture reveals the extraction and time efficiency of the proposed QPSONN architecture.
Handbook of Research on Machine Learning Innovations and Trends pp 1-39; https://doi.org/10.4018/978-1-4666-8291-7.ch001

Abstract:
In this chapter, an exhaustive survey of quantum behaved techniques on swarm intelligent is presented. The techniques have been categorized into different classes, and in conclusion, a comparison is made according to the benefits of the approaches taken for review. The above-mentioned techniques are classified based on the information they exploit, for instance, neural network related, meta-heuristic and evolutionary algorithm related, and other distinguished approaches are considered. Neural Network-Based Approaches exhibit a few brain-like activities, which are programmatically complicated, for instance, learning, optimization, etc. Meta-Heuristic Approaches update solution generation-wise for optimization, and the approaches differ based on the problem definition.
Sandip Dey, Siddhartha Bhattacharyya,
2014 International Conference on Computational Intelligence and Communication Networks pp 242-246; https://doi.org/10.1109/cicn.2014.63

Abstract:
In this paper, two quantum behaved multi-objective optimization techniques, based on Binary Particle Swarm Optimization and Ant Colony Optimization, have been introduced. The proposed approaches are used to search optimal threshold values of gray scale images, by optimizing the non-dominated solutions using Li's method as objective function. These approaches coalesce the meta-heuristic algorithms with the intrinsic features of quantum theory to make the techniques more effective. The best fitness values, the set of optimal thresholds and the computation times at different level of thresholding have been reported both for the proposed techniques and their equivalent classical counterparts. The superiority of the techniques presented in this paper, are established in terms of computational time. Thereafter, the mean fitness and the standard deviation of the objective values prove that the proposed techniques are more effectual of than others. Finally, the performance of each technique has been evaluated by determining the PSNR values of the test images. It was found that the proposed techniques have better PSNR values as compared to their corresponding components. Hence, quality of thresholding is established in favour of the proposed techniques.
C. Mala, M. Sridevi
2014 17th International Conference on Network-Based Information Systems pp 23-30; https://doi.org/10.1109/nbis.2014.15

Abstract:
Edge detection is an important process in image segmentation, object recognition, template matching, etc. It computes gradients in both horizontal and vertical directions of the image at each pixel position to find the image boundaries. The conventional edge detectors take significant time to detect the edges in the image. To reduce the computational time, this paper proposes parallel algorithms for edge detection with Sobel, Prewitt and Robert first order derivatives using a Shared Memory - Single Instruction Multiple Data (SM - SIMD) parallel architecture. From the experimental results, it is inferred that the proposed parallel algorithms for edge detection are faster than the conventional methods.
D. Amutha Devi, K. Muthukannan
2014 IEEE International Conference on Advanced Communications, Control and Computing Technologies pp 1374-1378; https://doi.org/10.1109/icaccct.2014.7019325

Abstract:
For any automated image analysis process, the segmentation is an important task because all subsequent tasks in image processing heavily rely on the quality of image segmentation. It determines the eventual success or failure of the analysis. The problem in image segmentation occurs when an image has a varying gray level background. There are several algorithms and methods are available for image segmentation, but there is a need to develop a unique method for it. In this paper, some of the image segmentation algorithms are compared to segment the diseased portion of rice leaves.
Sandip Dey, Siddhartha Bhattacharyya,
2014 International Conference on Computing for Sustainable Global Development (INDIACom) pp 311-316; https://doi.org/10.1109/indiacom.2014.6828150

Abstract:
In this article, a Quantum Inspired Tabu Search for Multi-level thresholding for Colour Image has been developed to boost the possible effectiveness than that of its classical counterpart. The proposed algorithm has been applied to two true colour images to determine optimal threshold values at different levels using Otsu's method as an objective function. The features of quantum mechanics are coupled with the basic constitution of a popular meta-heuristic algorithm, called tabu search to form the quantum inspired meta-heuristic algorithm. Between the participating algorithms, the proposed algorithm takes least time for execution. The usefulness of the proposed method is established in context of exactitude, resilience and computational time over its respective conventional method. In addition, a popular test, called one-tailed t-test, used for statistical measurement, demonstrates the efficiency of the proposed algorithm.
, , Susanta Chakraborty
Handbook of Research on Machine Learning Innovations and Trends pp 19-50; https://doi.org/10.4018/978-1-4666-4936-1.ch002

Abstract:
The optimized class responses from the image content has been applied to generate the optimized version of MUSIG (OptiMUSIG) activation function for a multilayer self organizing neural network architecture to effectively segment multilevel gray level intensity images. This chapter depicts the parallel version of the OptiMUSIG (ParaOptiMUSIG) activation function with the optimized class responses for the individual features with a parallel self-organizing neural network architecture to segment true color images. A genetic algorithm-based optimization technique has been employed to yield the optimized class responses in parallel. Comparison of the proposed method with the existing non-optimized method is applied on two real life true color images and is demonstrated with the help of three standard objective functions as they are employed to measure the quality of the segmented images. Results evolved by the ParaOptiMUSIG activation function are superior enough in comparison with the conventional nonoptimized MUSIG activation applied separately on the color gamut.
Sandip Dey, Siddhartha Bhattacharyya,
2013 Annual IEEE India Conference (INDICON) pp 1-6; https://doi.org/10.1109/indcon.2013.6726024

Abstract:
In this article, the particle swarm optimization and differential evolution algorithms inspired by the intrinsic principles of quantum mechanics are presented. These quantum versions of meta-heuristic algorithms, namely quantum inspired particle swarm optimization and quantum inspired differential evolution for multi-level thresholding have been designed to find optimal thresholds of colour images at different levels by exploiting Kapur's entropy as an objective function. The average fitness and the standard deviation of the fitness values are reported. The test results over two test images at different levels certify the efficacy of the proposed methods with reference to precision, computational time, and durability over their classical counterparts. At last, a statistical measure, t-test has been performed among the four methods (two quantum methods and two classical methods) taking two methods in a single grasp to ascertain the supremacy of the results.
Alireza Rezvanifar, Mohammadali Khosravifard
IEEE Transactions on Image Processing, Volume 23, pp 635-644; https://doi.org/10.1109/tip.2013.2289984

Abstract:
Applying a fast over-segmentation algorithm to image and working on a region-based graph (instead of the pixel-based graph) is an efficient approach to reduce the computational complexity of graph-based image segmentation methods. Nevertheless, some undesirable effects may arise if the conventional cost functions, such as Ncut, AverageCut, and MinCut, are employed for partitioning the region-based graph. This is because these cost functions are generally tailored to pixel-based graphs. In order to resolve this problem, we first introduce a new class of cost functions (containing Ncut and AverageCut) for graph partitioning whose corresponding suboptimal solution can be efficiently computed by solving a generalized eigenvalue problem. Then, among these cost functions, we propose one that considers the size of regions in the partitioning procedure. By simulation, the performance of the proposed cost function is quantitatively compared with that of the Ncut and AverageCut.
A. Devi Priya, G. V. S. Kumar
2013 International Conference on Information Communication and Embedded Systems (ICICES) pp 992-997; https://doi.org/10.1109/icices.2013.6508384

Abstract:
This paper presents a scheme for segmentation of texture images combining M-band wavelet transform and Fuzzy C-Means. M-band wavelet transform yields a large number of sub images which enhances the performance. M-Band Wavelet transform decomposes an image in to M×M channels. Different combinations of these band pass sections produce various scales and orientations in frequency plane, hence it produces a sixteen sub-band images. These features are subjected to Fuzzy C-Means clustering technique for segmentation. The advantage of FCM is that it does not require a priori knowledge to segment a region. This new combined algorithm produces good segmentation results by applying FCM for M-Band Wavelet extracted features.
Siddhartha Bhattacharyya, Sandip Dey
2011 International Conference on Computational Intelligence and Communication Networks pp 121-125; https://doi.org/10.1109/cicn.2011.24

Abstract:
A genetic algorithm inspired by the inherent features of parallelism and time discreteness exhibited by quantum mechanical systems, is presented in this article. The predominant interference operator in the proposed quantum inspired genetic algorithm (QIGA) is influenced by time averages of different random chaotic map models derived from the randomness of quantum mechanical systems. Subsequently, QIGA uses quantum inspired crossover and mutation on the trial solutions, followed by a quantum measurement on the intermediate states, to derive sought results. Application of QIGA to determine optimum threshold intensities is demonstrated on two real life gray level images. The efficacy of QIGA is adjudged w.r.t. a convex combination of two fuzzy thresholding evaluation metrics in a multiple criterion scenario. Comparative study of its performance with the classical counterpart indicates encouraging avenues.
Shunyong Zhou, Wenling Xie, Cuixia Guo, Bo Hu
2011 International Conference on Multimedia Technology pp 3810-3813; https://doi.org/10.1109/icmt.2011.6002920

Abstract:
A new color image segmentation algorithm based on histogram, FCM clustering, and region merging is proposed in this paper. First, the RGB space is transformed to HSV space, and the image is divided into non-singular points and singular points in accordance with the saturation. Second, characteristics of the image pixel are mapped to the one-dimensional histogram, we can determine the number of the cluster and the initial cluster center thought peaks selection algorithm, non-singular points and singular points are separately clustering by FCM,. Finally, we merger regions by image spatial information to eliminate the scattered small area after clustering, which overcomes the over segmentation problem in FCM, and increases the ability of anti noise. Experimental results show that this method not only can make the partition consistent with the human visual psychology, but also overcome the singularity of HSV space, and significantly reduce computational complexity and greatly improve the speed of the algorithm, realize automatically dividing images without manual intervention.
Page of 1
Articles per Page
by
Show export options
  Select all
Back to Top Top