Search

Facial expressions, which are important social communication tools in our daily life, provide important information about the mental state of people. Research is being done to obtain this information accurately. The importance of these researchs in the field of human-computer interaction is increasing. Many methods have been used for the recognition of universal facial expressions such as neutral, happiness, surprise, sadness, anger, disgust, and fear by intelligent systems with high accuracy. Emotion recognition is an example of difficult classification due to factors such as ambient light, age, race, gender, and facial position. In this article, a 3-stage system is proposed for emotion detection from facial images. In the first stage, the CNN-based network is trained with the Fer+ dataset. The Binary Particle Swarm Optimization algorithm is applied for feature selection to the feature vector in the fully connected layer of the CNN network trained in the second stage. Selected features are classified by Support Vector Machine. The performance of the proposed system has been tested with the Fer+ dataset. As a result of the test, 85.74% accuracy was measured. The results show that the combination of BPSO and SVM contributes to the classification accuracy and speed of the FER+ dataset.

Plant diseases and pests cause significant losses in agriculture, with economic, ecological and social implications. Therefore, early detection of plant diseases and pests via automated methods are very important. Recent machine learning-based studies have become popular in the solution of agricultural problems such as plant diseases. In this work, we present two classification models based on deep feature extraction from pre-trained convolutional neural networks. In the proposed models, we fine-tune and combine six state-of-the-art convolutional neural networks and evaluate them on the given problem both individually and as an ensemble. Finally, the performances of different combinations based on the proposed models are calculated using a support vector machine (SVM) classifier. In order to verify the validity of the proposed model, we collected Turkey-PlantDataset, consisting of unconstrained photographs of 15 kinds of disease and pest images observed in Turkey. According to the obtained performance results, the accuracy scores are calculated as 97.56% using the majority voting ensemble model and 96.83% using the early fusion ensemble model. The results demonstrate that the proposed models reach or exceed state-of-the-art results for this problem.

Melanoma is considered the deadliest form of skin cancer, and the number of cases is increasing day by day. The early diagnosis of melanoma is critical, as it significantly increases the patient’s chance of survival. However, distinguishing melanoma from other skin lesion types by the physician can be a complicated process due to the diversity of its structural and textural features. Numerous computer-aided diagnosis (CAD) systems have been developed to assist the physician in detecting melanoma during recent years. The segmentation is a critical step for CAD systems, as it directly contributes to the performance of both feature extraction and classification steps. The optimization of the hyperparameters of deep learning methods is a challenging research topic. In this paper, the Bayesian optimized SegNet approach is proposed for precise skin lesion segmentation. The proposed method is obtained competitive results with the latest skin lesion segmentation methods. The hyperparameters optimized SegNet has achieved the best results with the average Jaccard Index of 84.9 on ISBI2016 and 74.5 on ISBI2017 dataset. Experimental results indicate the validity of Bayesian optimized SegNet. In this study, it has been observed that the bayesian hyperparameter optimization in the SegNet, which is the latest deep learning architecture, increased the segmentation performance of the SegNet by 16% in the ISBI2016 dataset and by 7% in the ISBI2017 dataset.

Image processing-based fire/flame detection has become popular in recent years. In this paper, a novel fire/flame detection system based on a new conversion matrix and artificial bee colony algorithm was presented. Flame and non-flame image pixel values were combined to have a new feature matrix. A conversion matrix was generated randomly. The conversion matrix was multiplied by the feature matrix. The error of this multiplication result was calculated using the K-means clustering algorithm. The conversion matrix was updated until getting desired performance using artificial bee colony algorithm. At the end of the updating process, updated conversion matrix was multiplied with all images in the dataset to move all images to new color space. The final images were converted into binary images. Otsu method was used to get binary images. These binary images were compared with the corresponding ground truth images in the dataset. The aim of this comparison is to calculate the similarity ratio of the two images. This ratio shows the extent to which the original image features are preserved. A forest fire dataset was used which has 500 forest fire images. It is publicly available and called as Corsican Fire Database. Jaccard and Dice similarity measure parameters were used to evaluate the proposed system performance and compared with other similar study such as particle swarm optimization. Evaluated mean Jaccard index value was 0.76, and mean Dice index value was 0.85. This evaluation was made for 500 images. These results provide that this system can be used in fire/flame detection systems.

In this paper, we proposed Multi-model LSTM-based Pre-trained Convolutional Neural Networks (MLP-CNNs) as an ensemble majority voting classifier for the detection of plant diseases and pests. The proposed hybrid model is based on the combination of LSTM network with pre-trained CNN models. Specifically, in transfer learning, we adopted deep feature extraction from various fully connected layers of these pre-trained deep models. AlexNet, GoogleNet and DenseNet201 models are used in this work for feature extraction. The extracted deep features are then fed into the LSTM layer in order to construct a robust hybrid model for apple disease and pest detection. Later, the output predictions of three LSTM layers determined the class labels of the input images by majority voting classifier. In addition, we use an automatic scheme for determining the best choice of the network parameters of the LSTM layer. The experiments are carried out using data consisting of real-time apple disease and pest images from Turkey and the accuracy rates are calculated for performance evaluation. The experimental results show that by using the proposed ensemble combination structure, the results are comparable to, or better than, the pre-trained deep architectures.

Plants play a crucial role in the lives of all living things. A risk of extinction exists for many plants, hence many botanists and scientists are working in order to protect plants and plant diversity. Plant identification is the most important part of studies carried out for this purpose. In order to identify plants more accurately, different approaches have been used in the studies to date. One of these approaches is plant identification through leaf recognition, and is the basis of many conducted studies. It can be used for automatic plant recognition in the area of botany, the food sector, industry, medicine, and in many more areas too. In this study, image processing based on feature extraction methods such as color features, vein features, Fourier Descriptors (FD), and Gray-Level Co-occurrence Matrix (GLCM) methods are used. This study suggests the use of features extracted from leaves divided into two or four parts, instead of extracting for the whole leaf. Both the individual and combined performances of each feature extraction method are calculated by Extreme Learning Machines (ELM) classifier. The suggested approach has been applied to the Flavia leaf dataset. 10-fold cross-validation was used to evaluate the accuracy of the proposed method, which was then compared and tabulated with methods from other studies. The evaluated accuracy of the proposed method on the Flavia leaf dataset was calculated as 99.10%.

The timely and accurate diagnosis of plant diseases plays an important role in preventing the loss of productivity and loss or reduced quantity of agricultural products. In order to solve such problems, methods based on machine learning can be used. In recent years, deep learning, which is especially widely used in image processing, offers many new applications related to precision agriculture. In this study, we evaluated the performance results using different approaches of nine powerful architectures of deep neural networks for plant disease detection. Transfer learning and deep feature extraction methods are used, which adapt these deep learning models to the problem at hand. The utilized pretrained deep models are considered in the presented work for feature extraction and for further fine-tuning. The obtained features using deep feature extraction are then classified by support vector machine (SVM), extreme learning machine (ELM), and K-nearest neighbor (KNN) methods. The experiments are carried out using data consisting of real disease and pest images from Turkey. The accuracy, sensitivity, specificity, and F1-score are all calculated for performance evaluation. The evaluation results show that deep feature extraction and SVM/ELM classification produced better results than transfer learning. In addition, the fc6 layers of the AlexNet, VGG16, and VGG19 models produced better accuracy scores when compared to the other layers.

Hyperspectral images are images containing rich spectral and spatial information widely used in remote sensing applications. The development of deep learning techniques has had a significant impact on the classification of hyperspectral images. Different Convolutional Neural Network architectures have been used in many hyperspectral image analysis studies. However, the high dimensions of the hyperspectral images increased the computational complexity. For this reason, dimensionality reduction has been used in the preprocessing stage in many studies. Another difficulty encountered in hyperspectral image classification studies is the need to consider both spectral and spatial features. When deep spatial and spectral features are to be extracted, problems such as loss of gradient properties and degradation due to increased depth arise. In this study, the 3D convolutional neural network (CNN) based ResNet50 method is proposed to solve these problems encountered in hyperspectral studies and to extract sufficient spatial spectral properties from the network. Principal Component Analysis (PCA) was used to reduce spectral band excess. The proposed method has been applied to Pavia University and Salinas data sets. Overall accuracy, average accuracy and kappa values were used to measure the performance of the method. Calculated overall accuracy, average accuracy, and kappa values are 99.99% for the Pavia University data set, and while the overall accuracy and kappa values were 99.99% for the Salinas data set, the average accuracy value was 99.98%.

Surface defects detection in products used in industry such as steel, fabric and marble is very important in terms of increasing product quality and preventing financial losses. However, automatic surface defects detection is a very difficult problem due to the complexity and diversity of surface defects. In this study, U-net based VGG16-Unet and Resnet34- Unet network models are proposed for Surface defects detection. The proposed model used spatial features in the first layers together with deep semantic features. In the proposed network models, the trained weights of the VGG16 and Resnet34 network architectures were used for the input parameters of the Unet architecture. In experimental studies, the highest F1-score value for MT and AITEX data sets was obtained as 91.07% and 94.67%, respectively, with the proposed Resnet34-Unet model. According to the results, it was observed that the defective areas showing similarity with the background were successfully separated by using the proposed model.

Background Plants have an important place in the life of all living things. Today, there is a risk of extinction for many plant species due to climate change and its environmental impact. Therefore, researchers have conducted various studies with the aim of protecting the diversity of the planet’s plant life. Generally, research in this area is aimed at determining plant species and diseases, with works predominantly based on plant images. Advances in deep learning techniques have provided very successful results in this field, and have become widely used in research studies to identify plant species. Methods In this paper, a Multi-Division Convolutional Neural Network (MD-CNN)-based plant recognition system was developed in order to address an agricultural problem related to the classification of plant species. In the proposed system, we divide plant images into equal nxn-sized pieces, and then deep features are extracted for each piece using a Convolutional Neural Network (CNN). For each part of the obtained deep features, effective features are selected using the Principal Component Analysis (PCA) algorithm. Finally, the obtained effective features are combined and classification conducted using the Support Vector Machine (SVM) method. Results In order to test the performance of the proposed deep-based system, eight different plant datasets were used: Flavia, Swedish, ICL, Foliage, Folio, Flower17, Flower102, and LeafSnap. According to the results of these experimental studies, 100% accuracy scores were achieved for the Flavia, Swedish, and Folio datasets, whilst the ICL, Foliage, Flower17, Flower102, and LeafSnap datasets achieved results of 99.77%, 99.93%, 97.87%, 98.03%, and 94.38%, respectively.

Brain tumors are formed due to the abnormal growth of the cells which become proliferated in an uncontrollable way [1]..

A lot of important disease information can be accessed by performing retinal blood vessel analysis on fundus images. Diabetic retinopathy is one of the diseases understood by retinal blood vessel analysis. If this disease is detected at an early stage, vision loss can be prevented. In this paper, a method that performs retinal blood vessel analysis with classical methods is proposed. In this proposed system, pixel-based feature extraction is performed. Five different feature groups are used for feature extraction. These feature groups are edge detection, morphological, statistical, gradient, and Hessian matrix. An 18-D feature vector is created for each pixel. This feature vector is given to the artificial neural network for training. Using test images, the system is tested on two publicly available datasets. Sensitivity, Specificity, and Accuracy performance measures were used as success measures. The similarity index between the segmented image and the ground truth is measure using Dice and Jaccard. The accuracy of the system was measured as 96.18% for DRIVE and 94.56% for STARE, respectively. Experimental results show that the proposed algorithm achieves satisfactory results. This method can be used as an automated retinal blood vessel segmenting system.

Fabric quality control is one of the most important phases of production in order to ensure high-quality standards in the fabric production sector. For this reason, the development of successful automatic quality control systems has been a very important research subject. In this study, we propose a Multiple Pooling and Filter approach based on a Deep Neural Network (MPF-DNN) for the classification of texture defects. This model consists of three basic stages: preprocessing, feature extraction, and classification. In the preprocessing stage, the texture images were first divided into n × n equal parts. Then, median filtering and pooling processes were applied to each piece prior to performing image merging. In the proposed pre-treatment stage, it is aimed to clarify fabric errors and increase performance. For the feature extraction stage, deep features were extracted from the texture images using the pretrained ResNet101 model based on the transfer learning approach. Finally, classification and testing procedures were conducted on the obtained deep-effective properties using the SVM method. The multiclass TILDA dataset was used in order to test the proposed model. In experimental work, the MPF-DNN model for all four classes achieved a significant overall accuracy score of 95.82%. In the results obtained from extensive experimental studies, it was observed that the proposed MPF-DNN model was more successful than previous studies that used pretrained deep architectures.

The feature selection can be defined as the selection of the best subset to represent the data set, that is, the removal of unnecessary data that does not affect the result. The efficiency and accuracy of the system can be increased by decreasing the size and the feature selection in classification applications. In this study, text classification was applied by using “20 news group” data published by Reuters news agency. The pre-processed news data were converted into vectors using the Doc2Vec method and a data set was created. This data set is classified by the Maximum Entropy Classification method. Afterwards, a subset of data sets was created by using the Mutual Information Method for the feature selection. Reclassification was performed with the resulting data set and the results were compared according to the performance rates. While the success of the system with 600 features was (0.9285) before the feature selection, (0.9285), then, the performance rates of the 200, 100, 50, 20 models were obtained as (0.9454, 0.9426, 0.9407, 0.9123), respectively. When the results were examined, the success of the 50-featured model was higher than the 600-featured model initially created.

Data, which is the new trend of 21st century, has made difficult to extract meaningful information from itself because of its increasing volume and variety Big data platforms and tools to analyze the size and type of data that cannot be processed by traditional methods have brought new perspectives to data analytics. Big Data are gathered from non-traditional sources such as wireless sensors, blogs, social media, emails etc. In addition, companies can make better insights of large data sets by using large data analysis mechanisms. In this study, the definition of big data, its relationship with other technologies, technologies used in big data area and general information about analysis techniques are given.

Feature selection can be defined as the selection of the best subset to represent the data set in machine learning applications, in other words extraction of the unnecessary data that has no effect on the result. In classification problems efficiency and accuracy of the system can be increased when the dimension is reduced by feature selection. In this study, text classifying application is performed by using the data set of “20 News Group” released in Reuters News Agent. The pre-processed news data were converted to vectors by using Doc2Vec method and the data set was created and classified by Naive Bayes method. Subsequently, a subset of the data set was formed by using heuristic methods that were inspired by nature (Whale and Gray Wolf Optimization Algorithms) and Chi-square method for feature selection. Then the reclassification was applied and the results were compared. While the success of the system with 600 features before the feature selection is 0.9214, the performance ratio of the 100 featured models created later is figured higher (0.94095 - 0.93833- 0.93619).

Honey bees are one of the most important pollinators for a wide range of products in the food chain. Today, with the help of developing technology, observing bees healthy controls is a very important field of study. In this study, the images taken in the natural environment of the bees were processed with Convolutional Neural Network (CNN) architecture and the health status of the bees were classified. The results obtained were promising for the studies to be carried out in this area. In addition, the structure of the CNN architectures was studied and the CNN architectures with different types and number of layers were compared with each other. As a result of the comparison, using the ideal number of convolution layers instead of using a great number of convolution layers for CNN architectures, increases the success. In addition, the use of normalization layers that serve as supporters in CNN architectures has been found to be very important for increasing success. In this study, 5 different CNN architectures were developed and the classification results obtained with these architectures were analyzed. Among the architectures developed, KM_1 network architecture has achieved the best results with a success rate of 92,42.

Recently, Convolutional Neural Networks (CNN), which is used in the solution of many image processing problems, has been used successfully for many problems in the agricultural field . In this study, for classification of plant species is proposed an approach based on the combination of deep architectures. Deep features were extracted from the plant leaves using the fc6 layer of the previously trained AlexNet and VGG16 models. Then, the reduction of the number of deep features by using the Principal Component Analysis (PCA) method was done quickly and the best distinguishing features were obtained. Finally, the classification performances were calculated using the K-Nearest Neighbor (KNN) method. Flavia and Swedish plant leaf data sets were used to test the proposed system. According to the experimental results, the accuracy scores for Flavia and Swedish data sets was obtained as 99.42% and 99.64%, respectively.

Thanks to the rapid increase in technology and electronic communications, e-mail has become a serious communication tool. In many applications such as business correspondence, reminders, academic notices, web page memberships, e-mail is used as primary way of communication. If we ignore spam e-mails, there remain hundreds of e-mails received every day. In order to determine the importance of received e-mails, the subject or content of each e-mail must be checked. In this study we proposed an unsupervised system to classify received e-mails. Received e-mails' coordinates are determined by a method of natural language processing called as Word2Vec algorithm. According to the similarities, processed data are grouped by k-means algorithm with an unsupervised training model. In this study, 10517 e-mails were used in training. The success of the system is tested on a test group of 200 e-mails. In the test phase M3 model (window size 3, min. Word frequency 10, Gram skip) consolidated the highest success (91%). Obtained results are evaluated in section VI.

In recent years, deep learning, which is widely used in machine learning and computer vision, offers many new solutions, especially for agricultural problems. In this study, an approach based on the combination of Convolutional Neural Networks (CNN) and Color-Local Binary Pattern (C-LBP) method is recommended for the determination of plant species. Deep features have been obtained from the fc6 layer of the AlexNet model, a pre-trained ESA architecture. Then, LBP method is applied to each channel of color images (R, G, B). Finally, the deep features and LBP features from each color channel were combined and classified by Support Vector Machine (SVM). To test the accuracy of the proposed approach, ICL and Folio leaf data sets commonly used in the literature have been used. According to this results, accuracy rates of 98.50% and 99.48% were calculated for ICL and Folio data sets, respectively. The experimental results indicate that the proposed model achieves better accuracy compared to previous studies.

The histograms of oriented gradients (HOG) and co-occurrence HOG (CoHOG) algorithms are simple and intuitive descriptors. However, the HOG and CoHOG algorithms based on gradient computation still have some shortcomings: they ignore meaningful textural properties and are unstable to noise. In this paper, two new efficient HOG and CoHOG methods are proposed. The proposed algorithms are based on the Gaussian derivative filters, and the feature vectors are obtained by means of principal curvatures. The feature vectors are rotation invariant by means of the rotation invariance characteristic of principal curvatures (i.e. eigenvalues). The experimental results on the CUReT, KTH-TIPS, KTH-TIPS2-a, UIUC, Brodatz album, Kylberg and Xu datasets confirm that the developed algorithms have higher classification rates than state-of-the-art texture classification methods. The classification results also demonstrate that the developed algorithms are more stable to noise and rotation than the original HOG and CoHOG algorithms. Two continuous rotation invariant descriptors are proposed.The proposed descriptors are based on principal curvatures which are rotation invariant.The experimental results show the power of the methods particularly in extremely noisy conditions.Our approaches give high classification performance on seven texture databases.

Exercise is a key component for prevention and treatment of type 2 diabetes. However, diabetes complications affect exercise habits. Computerized clinical decision support systems (CCDSSs) may help specialists improve their decision-making abilities in the management of diseases. We hypothesized that patients' diabetic neuropathy, neuropathic pain, and kinesiophobia will quickly be identified in the early stages by using the designed CCDSSs. It is thought that such systems will help in planning exercise programs for patients with diabetes and in maintaining the appropriate programs. Based on our hypothesis, we conclude that CCDSSs will also be effective in managing complications and movement dysfunctions occurring in the musculoskeletal system.

2019 International Conference on Artificial Intelligence and Data Processing Symposium, IDAP 2019 -- 21 September 2019 through 22 September 2019 -- -- 153040, Fabric defect detection is one of the most important areas for quality control of products in the textile industry. Many different studies have developed methods for this problem. In this study, an automatic thresholding method developed with entropy has been proposed. Due to the low cost of calculation, the proposed automatic thresholding method will be very suitable for real-Time applications. In this study, automatic thresholding method which is supported by 4 different entropy method was compared with otsu method which is one of automatic thresholding methods. Various tests have been made on different fabric types for comparisons. As a result of experimental studies, successful results of automatic thresholding methods supported with entropy were obtained for fabric defect detection. Renyi entropy method was the most successful result among the proposed methods. © 2019 IEEE.

Usage of color information of object is one of the most popular image processing based object detection method. In this work, a method has been proposed for separating color pixels of target objects from other object color pixels in the image. A fire/flame dataset was used for the target colored object. The proposed method consists of three stage. In the first stage, the flame and non-flame color pixels, taken from this dataset, are combined in a feature matrix. In the second stage, the feature matrix is subjected to linear multiplication with a 3*3 conversion matrix whose initial values are randomly generated. The error values of the linear multiplication result are calculated by the K-means algorithm and then these error values are optimized by the Biogeography Based Optimization (BBO) algorithm. In the last stage, the optimal conversion matrix is applied on the images and the color distribution of the target color pixels in three dimensional space is examined.

In recent years, deep learning widely used in image processing field, has introduced many new applications related to the agricultural field. In this study, for apricot disease detection were used deep learning models such as AlexNet, Vgg16, and Vgg19 based on pre-trained deep Convolutional Neural Networks (CNN). The deep attributes obtained from these models are classified by K-Nearest Neighbour (KNN) method. To calculate the performance of the proposed methods was applied 10- fold cross-validation test. The dataset consists of 960 images including healthy and diseased apricot images. According to the obtained results, the highest accuracy was obtained as 94.8% by using Vgg16 model.

Four highly discriminative and continuous rotation invariant methods are proposed.We use the Hessian matrix and Gaussian derivative filters.Verified on the CUReT, KTH-TIPS, KTH-TIPS2-a, UIUC and Brodatz texture datasets. Extracting rotation invariant features is a valuable technique for the effective classification of rotation invariant texture. The Histograms of Oriented Gradients (HOG) algorithm has been proved to be theoretically simple, and has been applied in many areas. Also, the co-occurrence HOG (CoHOG) algorithm provides a unified description including both statistical and differential properties of a texture patch. However, HOG and CoHOG have some shortcomings: they discard some important texture information and are not invariant to rotation. In this paper, based on the original HOG and CoHOG algorithms, four novel feature extraction methods are proposed. The first method uses Gaussian derivative filters named GDF-HOG. The second and the third methods use eigenvalues of the Hessian matrix named Eig(Hess)-HOG and Eig(Hess)-CoHOG, respectively. The fourth method exploits the Gaussian and means curvatures to calculate curvatures of the image surface named GM-CoHOG. We have empirically shown that the proposed novel extended HOG and CoHOG methods provide useful information for rotation invariance. The classification results are compared with original HOG and CoHOG algorithms methods on the CUReT, KTH-TIPS, KTH-TIPS2-a and UIUC datasets show that proposed four methods achieve best classification result on all datasets. In addition, we make a comparison with several well-known descriptors. The experiments of rotation invariant analysis are carried out on the Brodatz dataset, and promising results are obtained from those experiments.

Hilbert Transform (HT) and Multi Wavelet Transform (MWT) has been used to recognize the same frequency harmonics that occur in the brain with the Steady State Visual Evoked Potentials(SSVEP). In this study, harmonics of certain frequencies in brain are used which are detected by SSVEP and visual stimulus potentials to be used in Robot Arm Control. This stimulus has been made using shapes of box that oscillated at certain frequencies. The signal components were clustered according to the same direction and stimulus frequency on the data set for the desired work, task or movement. These signals were processed by the band pass filters at 5-30 Hz then HD process were applied. The filtered signals classified by Neural Network and Cubic-Support Vector Machine after MWT analysis were applied to these. Evaluated average success rate is over 90 %. Finally, the test brain signals recorded for 3 tasks over the trained network have been successfully used for Robot Arm Control. The use of the proposed HD- MWT method is promising for the development of a real-time robot control with SSVEP-based BCI.

The delays in the detection of fire in fire detection systems continue to be a life threatening problem for living things. Techniques based on image processing have been developed in order to remove this problem and minimize the detection period. This study also focused on the smoke image that appeared before the flame at the time of the fire. Smoke detection can provide earlier notification than flame detection. In the first step of the proposed method, smoke zone was detected with YUV color space. After than the Gray Level Co-Occurrence Matrix (GLCM) was used to extract the features that represent the smoke images. At last, these features are used to classify the smoke and non-smoke space by using Support Vector Machines (SVM).

Plant recognition from their leaves has become a popular area in the machine learning and image processing. In this study 7 different types of apricot trees were determined and classified by using their leaves. At first leaves images were pre-processed. After than each image was scanned by 5×5 overlapping filter and median values of each filter process were recorded to represent the leaves. After than filtered each image was scanned by 2×2 overlapping filter and maximum values of each shifting step was recorded. The dimension of each image reduced to it' half. Histogram of these uniform patterns were evaluated. These features were applied as input to the Artificial Neural Network (ANN) and 7 types of apricot were classified with the accuracy is 98.6 %.

Image processing based systems take important place in systems used to detect fires in open spaces. Vision-based systems can detect fires in open spaces from distances and detect the fire at an early stage. In this study, a fire/flame detection method based on the color analysis of the fire image is presented. The proposed method consists of three steps. First, the image in RGB space is converted to the HSI color space. Then a color filter is applied to determine the fire/flame candidate zone. In the second stage, fake fire zones within the candidate zone identified as fire are eliminated. In this phase, image difference and gauss mixture model is used to recover the fake fire areas. In the third step, the result of the two methods is subjected to AND processing. The AND operation ensures to detect the exact flame zone. As a result, the proposed algorithm has been tested using fire video images. The highest calculated accuracy is 96%.

Electromyogram (EMG) signal can be defined as the electrical activity of muscles cells. It is commonly used in motion recognition, treatment of neuromuscular disorders and prosthetic hand control. In this study, classification of EMG signals obtained from 6 different hand shapes of holding object was proposed. At first Short Time Fourier Transform of the EMG signal were evaluated to obtain their Time-Frekans representation. After than these T-F images were segmented and their mean values were evaluated to reduce the dimension of the images. Local Receptive Fields based Extreme Learning Machines (ELM-LRF) used to classification of these hand shapes of holding object. Evaluated accuracy is 94.12 %.

Over the past 15 years, many feature extraction methods have been used and developed for the recognition of plant species. These methods have mostly been performed using separation operations from the background based on a pre-processing stage. However, the Local Binary Patterns (LBP) method, which provides high performance in object recognition, is used to obtain textural features from images without need for a pre-processing stage. In this paper, we propose different approaches based on LBP for the recognition of plant leaves using extracted texture features from plant leaves. While the original LBP converts color images to gray tones, the proposed methods are applied by using the R and G color channel of images. In addition, we evaluate the robustness of the proposed methods against noise such as salt & pepper and Gaussian. Later, the obtained features from the proposed methods were classified and tested using the Extreme Learning Machine (ELM) method. The experimental works were performed using various plant leaf datasets such as Flavia, Swedish, ICL, and Foliage. According to the obtained performance results, the calculated accuracy values for Flavia, Swedish, ICL and Foliage datasets were 98.94%, 99.46%, 83.71%, and 92.92%, respectively. The results demonstrate that the proposed method was more successful when compared to the original LBP, improved LBP methods, and other image descriptors for both noisy and noiseless images.

Texture can be defined as a local statistical pattern of texture primitives in observer's domain of interest. Texture classification aims to assign texture labels to unknown textures, according to training samples and classification rules. In this paper a novel method, which is an intelligent system for texture classification is introduced. It used a combination of genetic algorithm, discrete wavelet transform and neural network for optimum feature extraction from texture images. An algorithm called the intelligent system, which processes the pattern recognition approximation, is developed. We tested the proposed method with several texture images. The overall success rate is about 95%.

This paper presents an expert system based on wavelet decomposition and neural network for modeling and simulation of Chua's circuit which is used for chaos studies. The problems which arise in modeling Chua's circuit by neural networks are high structural complexity and slow and difficult training. With this proposed method a new solutions is produced to solve these problems. Wavelet decomposition is used for new useful feature extracting from input signal and neural network is used for modeling. Test results of proposed wavelet decomposition and neural network model are compared with test results of neural network model. Desired performance is provided by this new model. Test results showed that the suggested method can be used efficiently for modeling nonlinear dynamical systems.

In this paper, an intelligent wastewater treatment plant model is developed to predict the performance of a wastewater treatment plant (WWTP). The developed model is based on wavelet packet decomposition, entropy and neural network. The data used in this work were obtained from a WWTP in Malatya, Turkey. Daily records of these WWTP parameters over a year were obtained from the plant laboratory. Wavelet packet decomposition was used to reduce the input vectors dimensions of intelligent model. The suitable architecture of the neural network model is determined after several trial and error steps. Total suspended solid is one of the measures of overall plant performance so the developed model is used to predict the total suspended solid concentration in plant effluent. According to test results, the developed model performance is at desirable level. This model is an efficient and a robust tool to predict WWTP performance.

This paper presents a comparative study of implementation of feature extraction and classification algorithms based on discrete wavelet decompositions and Adaptive Network Based Fuzzy Inference System (ANFIS) for digital modulation recognition. Here, in first stage, 20 different feature extraction methods are generated by separately using Daubechies, Biorthogonal, Coiflets, Symlets wavelet families. In second stage, the performance comparison of these feature extraction methods is performed by using a new Expert Discrete Wavelet Adaptive Network Based Fuzzy Inference System (EDWANFIS). The digital modulated signals used in this experimental study are ASK8, FSK8, PSK8, QASK8. EDWANFIS structure consists of two parts. The first part is Discrete Wavelet Transform (DWT)-adaptive wavelet entropy and Adaptive Network Based Fuzzy Inference System for Automatic Digital Modulation Recognition (ADMR). The performance of this comparison system is evaluated by using total 800 digital modulated signals for each of these feature extraction methods. The performance comparison of these features extraction methods and the advantages and disadvantages of the methods are examined.

The physical process of oxygen transfer or oxygen absorption from the atmosphere acts to replenish the used oxygen, a process termed re-aeration or aeration. Aeration enhancement by macro-roughness is well-known in water treatment and one form is the aeration cascade. The macro-roughness of the steps significantly reduces the flow velocities and leads to flow aeration along the stepped cascade. In this paper, the aeration efficiency in stepped cascade aerators was modeled by using the Adaptive Network Based Fuzzy Inference System (ANFIS). The obtained model was tested with experimental data. Test results showed that ANFIS can be used to estimate the aeration efficiency in stepped cascade aerators.

In today's technology, computer assisted detection applications have managed to make great contributions to the field of medicine. Computer assisted detection systems aim to help radiologist about mass detection by using image processing systems. In this study, it's aimed to carry out mass detection process on the images of 3D brain MRI (Magnetic Resonance Imaging). The steps followed in this study are the stage of pre-processing the stage of segmentation, identification of the areas of interests and identification of tumor. As a result of processing's in the stages of preprocessing and segmentation, obtained areas of interests are labelled and attributes of these areas of interests are extracted during the stage of attributes extraction and in the last stage, the areas of interests are identified as whether they are mass or not according to these attributes. With this method applied on 845 number of magnetic resonance image sections belonging to 13 patients, it has been achieved classification success with 86.39%.

In this paper, to classify the grape tree species, the extracted features from leaf images are classified using a multi-class support vector machines. Feature extraction stage, the grape leafs are calculated by using 9 different features. Image processing stage involves gray tone dial, median filtering, contrast, thresh holding and morphological-logical processes. In the classification stage, the obtained properties with the help of multi-class support vector machines (MCSVM) is performed classification process. In the testing phase, by applying the different leaf images is calculated the performance of model. In this study, MATLAB software was used. At the end of the test was determined the total success rate of 90.7%.

In this paper, mean shift algorithm and adaptive Kalman filter have been both utilized to realize object tracking in video sequences. Mean shift algorithm cannot give good results when the position of the tracked object is changed rapidly between sequential frames or the tracked object is occluded. In this paper, the first position of the tracked object is predicted by Kalman filter then mean shift algorithm starts to seek the object in this position. Bhattacharyya coefficient which is obtained from mean shift algorithm, is used to instantly update Kalman filters error covariance matrix and determine whether object is occluded or not. Experimental results demonstrate that the proposed method has been more efficient technique as compared to standard mean shift algorithm in case of occlusion and fast object tracking.

Object detection is currently one of the most actively researched areas of computer vision, image processing and analysis. Image co-occurrence has shown significant performance on object detection task because it considers the characteristic of objects and spatial relationship between them simultaneously. CoHOG has achieved great success on different object detection tasks, especially human detection. Whereas, CoHOG is sensitive to noise and it does not consider gradient magnitude which significantly effects the object detection accuracy. To overcome these disadvantages the CoGMuLBP was proposed. CoGMuLBP uses a new statistical orientation assignment method based on uniform LBP instead of using the common gradient orientation. In this study, detection accuracies of CoGMuLBP and CoHOG are calculated on three different datasets with NN classifier. In addition, to evaluate the noise performance of the methods, gaussian noises were added to test images and performances were recalculated. Numerical experiments performed on three different datasets show that 1) CoGMuLBP has higher detection accuracy than CoHOG; 2) using uniform LBP based gradient orientation improves detection accuracy; and 3) CoGMuLBP is more robust to gaussian noise and illumination changes. These results provide the effectiveness of CoGMuLBP for object detection.