جستجوی مقالات مرتبط با کلیدواژه « پردازش تصویر » در نشریات گروه « برق »

Subject: 

Today, with the advancement of technologies to assist blind and visually impaired people, navigation systems are of great importance. As a result of emerging technologies in telecommunication and smartphones, these people can be helped. Identifying pedestrian and traffic lights is important to help pedestrians with visual impairments cross the intersection safely and securely. Background- researchers have studied the detection and identification of traffic lights in the assistive system or blind assist device. These researches can be divided into three main types: based on pattern matching, based on circular shape extraction, and based on color distribution.

In this research, an architecture based on mobile cloud computing is proposed, which can help blind pedestrians in crossing intersections. The architecture consists of three tiers: mobile phone, cloud, and supervision. The most important component is located on the mobile phone. It recognizes the color of pedestrian light by using image processing techniques. Spatial information (time and location) of the blind person is collected and held in a cloud storage database so that acquaintances can monitor him if needed. In order to detect the status of pedestrian lights, pictures of crossing streets with cameras will be captured. Using the features of color and morphology operations, the color of pedestrian lights is recognized and reported to the blind person. To this end, morphological operations are performed to eliminate small elements in the background and to restore the original size of the traffic light sign. Therefore, the operations of dilation, filling, and erosion are used.

We gathered a dataset including 280 photos of pedestrian lights (170 photos at day, 110 photos at night) in different illumination conditions (early day, noon, early night, night) and weather (sunny, cloudy, rainy). Matlab software and notebook system with Intel (R) Core (TM) i5 CPU and AMD Mobility Radeon HD 5100 graphics card were used to implement pedestrian traffic-light status detection. The scenario-based method is used to evaluate the system architecture and show that the proposed system can satisfy the investigated scenario. At last, the implementation results on taken images show excellent performance in detecting pedestrian lights with approximately 100% accuracy for the day and night.

The same size of the helix steps in a traveling wave tube has a direct relationship with the efficiency of the system. In this article, the goal is to build a device that measures helix steps automatically and with proper accuracy by using image processing. For this purpose, the proposed algorithm was presented along with innovations in the fields of noise removal and threshold determination for the automatic measurement of helix steps, and the prototype of the device was made. This device has the ability to work in two modes of manual and automatic measurement, and it has no domestic equivalent, and it is much cheaper in terms of cost and use in measuring helix steps than the sample made abroad. In automatic mode, this device is able to measure all helix steps at once. In the case of a similar sample made abroad the steps must be measured manually and separately. The selection of the beginning and end points of steps in this device is done adaptively, which increases the speed of measurement and increases the repeatability in measuring steps. In the sample made abroad, the selection of the start and end points of the steps is done by the user, which reduces the repeatability of the measurement. The obtained results show that the measurement accuracy of the manufactured device is acceptable compared to the sample made abroad.

Detection of salient objects is done with the aim of identifying and segmenting prominent objects or areas in an image. Fully Convolutional Networks (FCNs) have shown their advantages in salient object detection; however, many previous works have focused on the accuracy of the prominent area without paying attention to its edge. This paper focuses on the complementarity between edge information and salient object one and added an edge recognition module to explicitly model edge information to maintain salient object boundaries. Our proposed network is trying to improve these two tasks simultaneously. The presence of objects with different scales in related datasets is another problem in this area. It requires an appropriate cost function to deal with the imbalance problem between background and foreground in images. So, we have used the hybrid cost function in the training phase, which is not sensitive to the scale of objects and can better manage the problem of spatial coherence and uniformly highlight salient areas without additional parameters. A Comparison of the quantitative and qualitative results obtained by the proposed method with other advanced methods in six widely used protrusion detection datasets shows that the proposed method has a good performance and can quickly identify prominent areas. In particular, according to the quantitative results, our method gets the best result on three widely used test datasets in terms of F-measure and MAE criteria, demonstrating the proposed method's efficiency.

Sewage flow path is the main component of urban sewerage network infrastructure. Damage to sewers is less noticeable due to invisibility, and this failure to handle the damage leads to emergencies and unreasonable costs. These vital arteries need to be maintained and rebuilt during service for optimal performance in all dimensions. Nowadays, the methods of processing and classifying photos and videos taken by mobile videometer robots are widely used to inspect the sewer network. One of the successful algorithms in the field of image processing is the convolutional neural network algorithm, which is a subset of deep learning algorithm. In this paper, a convolutional neural network algorithm is used to classify images of sewer network damage and cases affecting the improvement, accuracy and performance of this algorithm. The images were obtained by a videometric robot from the sewer network. Results of using the proposed algorithm in the sewerage network, achieving 98% accuracy in classifying network faults and compared to other methods and also reducing the relatively low execution time of the proposed architecture (91 minutes) compared to other architectures valid ones are the same in deep learning on the same hardware platform. Also, in the future, the proposed algorithm will be used to analyze networks without the need for specialized personnel and also to control an automatic network videometry robot.

Fighting forest fires to avoid their potential dangers as well as protect natural resources is a challenge for researchers. The goal of this research is to identify the features of fire and smoke from the unmanned aerial vehicle (UAV) visual images for classification, object detection, and image segmentation. Because forests are highly complex and nonstructured environments, the use of the vision system is still having problems such as the analogues of flame characteristics to sunlight, plants, and animals, or the smoke blocking the images of the fire, which causes false alarms. The proposed method in this research is the use of convolutional neural networks (CNNs) as a deep learning method that can automatically extract or generate features in different layers. First, we collect data and increase them according to data augmentation methods, and then, the use of a 12-layer network for classification as well as transfer learning method for segmentation of images is proposed. The results show that the data augmentation method used due to resizing and processing the input images to the network to prevent the drastic reduction of the features in the original images and also the CNNs used can extract the fire and smoke features in the images well and finally detect and localize them.

Today, with the ever-increasing use of mobile devices, power consumption besides cost and performance has become an important parameter in the design of digital systems. Approximate computational blocks, with acceptable accuracy, have a positive impact on all three criteria of cost, power, and performance; approximate or inaccurate calculations has been accepted more than ever as a new approach to the design of digital systems. In this paper, an approximate adder is designed and the error rate is evaluated. Then, by using this adder, an approximate multiplier is also proposed. This multiplier is used in smoothing and sharpening application in the image processing applications. Using these computing blocks has led to saving 63% of dynamic power consumption, 70% static power saving, 25% decrease in design area and 33% improvement in latency. Also, the amount of error of system due to the use of approximate calculations is on average 33%, which does not have a negative effect on the function of the application in smoothing and sharpening.

In forensic dentistry, age is estimated using dental radiographs. Our goal is to automate these steps using image processing and pattern recognition techniques. With a dental radiograph, the contour is extracted and features such as apex, width and tooth length are determined, which are used to estimate age. Optimizing the resolution of radiographic images is an important step in contour extraction and age estimation. In this article, the aim is to improve the image resolution in order to extract the appropriate area and proper segmentation of the tooth, which makes it possible to estimate age better. In this model, due to the low resolution of radiographic images, in order to increase the accuracy of extracting the desired area of each tooth (ROI), the image resolution increases using spatial entropy based on the spatial distribution of pixel brightness, along with another increasing resolution method, like the Laplacian pyramids. Increasing the resolution of the image leads to the extraction of appropriate ROI and the removal of unwanted areas. The database used in this study is 154 adolescent panoramic radiographs, of which 73 are male and 81 are female. This database is prepared from Babol University of Medical Sciences. The results show that by using fixed tooth segmentation methods and only by applying the proposed effective method to improve image resolution, the extraction of appropriate ROI increased from 66% to 78% which shows a good improvement. The extracted ROI is then delivered to the segmented block and the contour extracted. After contour extraction, age is estimated. The age estimation using the proposed method is closer to the manual age estimate compared to the method that does not use the proposed algorithm to increase the image resolution.

Due to the growing increase of generated images via cameras and various instruments, image processing has found an important role in most of practical usages including medical, security and driving. However, most of the available models has no considerable performance and in some usages the amount of error is very effective. The main cause of this failure in most of available models is the distribution mismatch across the source and target domains. In fact, the made model has no generalization to test data with different properties and distribution compared to the source data, and its performance degrades dramatically to face with new data. In this paper, we propose a novel approach entitled Sparse coding and ADAptive classification (SADA) which is robust against data drift across domains. The proposed model reduces the distribution difference across domains via generating a common subspace between the source and target domains and increases the performance of model. Also, SADA reduces the distribution mismatch across domains via the selection of the source samples which are related to target samples. Moreover, SADA adapts the model parameters to build an adaptive model to encounter with data drift. Our variety of experiments demonstrate that the proposed approach outperforms all stat-of-the-art domain adaptation methods.

After the discovery of X-rays with the increasing use of digital imaging systems, medical image processing has become more important. Medical image processing helps specialists in diagnosing diseases. In addition to major digital techniques such as computed tomography (CT) or magnetic resonance imaging (MRI), analog imaging techniques such as endoscopy or radiography are now equipped with digital sensors. By processing images using different methods, the procedure applied to patients can be improved. Algorithms play a key role in noise filtering, segmentation, extraction, and characterization that diagnose diseases. MATLAB software and image processing toolboxes provide a wide range of advanced image processing functions and interactive tools for enhancing and analyzing digital images. In this article, using several algorithms designed in MATLAB, the quality of images is examined and a more appropriate algorithm is selected.

Domain adaptation can transfer knowledge from a training set (source domain) to a test set (target domain), promoting the performance of the model learned from the training set. In addition, sparse coding makes the learned model more succinct and easy to manipulate. However, the existence of the distribution mismatch across the source and target domains reduce the performance of model. In this paper, we propose an unsupervised domain adaptation model to minimize the prediction error of image classification. Sample reweighting is utilized to handle redundant and useless information of source data in the new representation. Moreover, the difference of the conditional distributions across the source and target domains is reduced along with the subspace alignment. Our proposed approach learns a sparse domain-invariant classifier in a latent subspace with preserving the structure of the input data. Extensive experiments demonstrate that our proposed approach shows 4.49% improvement in classification accuracy on real-world datasets compared to state-of-the-art machine learning and domain adaptation methods.

Image processing is a method to perform some operations on an image, in order to get an enhanced image or to extract some useful information from it. The conventional image processing algorithms cannot perform well in scenarios where the training images (source domain) that are used to learn the model have a different distribution with test images (target domain). In fact,the existence of conditional distribution difference across the source and target domains degrades the performance of model. Domain adaptation and transfer learning are promising solutions that aim to generalize a learning model across training and test data with different distributions. In this paper, we address the problem of unsupervised cross domain image processing in which no labels are available in test images. In fact, the proposed method transfers the source and target domains into a shared low dimensional FLDA-based subspace in an unsupervised manner. Our proposed method minimizes the conditional probability distribution difference of the source and target data via Bregman divergence. We provide a projection matrix to map the source and target data into a common subspace on which the between class scatter matrix is maximized and within class scatter matrix and cross domain distributions are minimized. Extensive experiments on 58 cross-domain image classification tasks over six public datasets reveal that our proposed method outperforms the state-of-the-art cross domain image processing approaches.

In this paper, a new method for image denoising based on incoherent dictionary learning and domain transfer technique is proposed. The idea of using sparse representation concept is one of the most interesting areas for researchers. The goal of sparse coding is to approximately model the input data as a weighted linear combination of a small number of basis vectors. Two characteristics should be considered in the dictionary learning process: Atom-data coherence and mutual coherence between dictionary atoms. The first one determines the dependency between the dictionary atoms and training data frames. This criterion value should be high. Another parameter expresses the dependency between atoms defined as the maximum absolute value of the cross-correlations between them. Higher coherence to the data class and lower mutual coherence between atoms result in a small approximation error in sparse coding procedure. In the proposed dictionary learning process, a coherence criterion is employed to yield over complete dictionaries with the incoherent atoms. The purpose of learning dictionary with low mutual coherence value is to reduce the approximation error of sparse representation in the denoising process and also decrease the computing time. We utilize the least angle regression with coherence criterion (LARC) algorithm for sparse representation based on atom-data coherence in the first step of dictionary learning process. LARC sparse coding is an optimized generalization of the least angle regression algorithm with stopping condition based on a residual coherence. This approach is based on setting a variable cardinality value. Using atom-data coherence measure as stopping criteria in the sparse coding process yields the capability of balancing between source confusion and source distortion. A high value for the cardinality parameter or too dense coding results in the source confusion since the number of dictionary atoms is more than what is required for a proper representation. Source degradation occurs when the sparse coding is done with low cardinality parameter or too sparse coding. Therefore, the number of required atoms will not be enough and data cannot be coded exactly over these atoms. Therefore, the setting procedure of cardinality parameter must be performed precisely. The problem of finding a dictionary with low mutual coherence between its normalized atoms can be obtained by considering the Gram matrix. The mutual coherence is described by the maximum absolute value of the off-diagonal elements of this matrix. If all off-diagonal elements are the same, a dictionary with minimum self-coherence value is obtained. Also, we take advantage of domain adaptation technique to transfer a learned dictionary to an adapted dictionary in the denoising process. The initial atoms set randomly and are updated based on the selected patches of input noisy image using the proposed alternating optimization algorithm. According to these issues, the fitness function in dictionary learning problem includes three main sections: The first term is related to the minimization of approximation error. The next items are the incoherence criterion of dictionary atoms. The last one includes a transformation of initial atoms according to some patches of the noisy input data in the test step. We use limited-memory BFGS algorithm as an iterative solution for regular minimization of our objective function involved different terms. The simulation results show that the proposed method leads to significantly better results in comparison with the earlier methods in this context and the traditional procedures.

Image processing is a method to perform some operations on an image, in order to get an enhanced image or to extract some useful information from it. The conventional image processing algorithms cannot perform well in scenarios where the training images (source domain) that are used to learn the model have a different distribution with test images (target domain). Also, many real world applications suffer from a limited number of training labeled data and therefore benefit from the related available labeled datasets to train the model. In this way, since there is the distribution difference across the source and target domains (domain shift problem), the learned classifier on the training set might perform poorly on the test set. Transfer learning and domain adaptation are two outstanding solutions to tackle this challenge by employing available datasets, even with significant difference in distribution and properties, to transfer the knowledge from a related domain to the target domain. The main assumption in domain shift problem is that the marginal or the conditional distribution of the source and the target data is different. Distribution adaptation explicitly minimizes predefined distance measures to reduce the difference in the marginal distribution, conditional distribution, or both. In this paper, we address a challenging scenario in which the source and target domains are different in marginal distributions, and the target images have no labeled data. Most prior works have explored two following learning strategies independently for adapting domains: feature matching and instance reweighting. In the instance reweighting approach, samples in the source data are weighted individually so that the distribution of the weighted source data is aligned to that of the target data. Then, a classifier is trained on the weighted source data. This approach can effectively eliminate unrelated source samples to the target data, but it would reduce the number of samples in adapted source data, which results in an increase in generalization errors of the trained classifier. Conversely, the feature-transform approach creates a feature map such that distributions of both datasets are aligned while both datasets are well distributed in the transformed feature space. In this paper, we show that both strategies are important and inevitable when the domain difference is substantially large. Our proposed using sample-oriented Domain Adaptation for Image Classification (DAIC) aims to reduce the domain difference by jointly matching the features and reweighting the instances across images in a principled dimensionality reduction procedure, and construct new feature representation that is invariant to both the distribution difference and the irrelevant instances. We extend the nonlinear Bregman divergence to measure the difference in marginal, and integrate it with Fisher’s linear discriminant analysis (FLDA) to construct feature representation that is effective and robust for substantial distribution difference. DAIC benefits pseudo labels of target data in an iterative manner to converge the model. We consider three types of cross-domain image classification data, which are widely used to evaluate the visual domain adaptation algorithms: object (Office+Caltech- 256), face (PIE) and digit (USPS, MNIST). We use all three datasets prepared by and construct 34 cross-domain problems. The Office-Caltech-256 dataset is a benchmark dataset for cross-domain object recognition tasks, which contains 10 overlapping categories from following four domains: Amazon (A), Webcam (W), DSLR (D) and Caltech256 (C). Therefore 4 × 3 = 12 cross domain adaptation tasks are constructed, namely A → W, ..., C → D. USPS (U) and MNIST (M) datasets are widely used in computer vision and pattern recognition tasks. We conduct two handwriting recognition tasks, i.e., usps-mnist and mnist-usps. PIE is a benchmark dataset for face detection task and has 41,368 face images of size 3232 from 68 individuals. The images were taken by 13 synchronized cameras and 21 flashes, under varying poses, illuminations, and expressions. PIE dataset consists five subsets depending on the different poses as follows: PIE1 (C05, left pose), PIE2 (C07, upward pose), PIE3 (C09, downward pose), PIE4 (C27, frontal pose), PIE5 (C29, right pose). Thus, we can construct 20 cross domain problems, i.e., P1 → P2, P1 → P3, ..., P5 → P4. We compare our proposed DAIC with two baseline machine learning methods, i.e., NN, Fisher linear discriminant analysis (FLDA) and nine state-of-the-art domain adaptation methods for image classification problems (TSL, DAM, TJM, FIDOS and LRSR). Due to these methods are considered as dimensionality reduction approaches, we train a classifier on the labeled training data (e.g., NN classifier), and then apply it on test data to predict the labels of the unlabeled target data. DAIC efficiently preserves and utilizes the specific information among the samples from different domains. The obtained results indicate that DAIC outperforms several state of-the-art adaptation methods even if the distribution difference is substantially large.

Nowadays automated early warning systems are essential in human life. One of these systems is fire detection which plays an important role in surveillance and security systems because the fire can spread quickly and cause great damage to an area. Traditional fire detection methods usually are based on smoke and temperature detectors (sensors). These methods cannot work properly in large space and out-door environments. They have high false alarm rates, and to cover the entire area, many smoke or temperature fire detectors are required, that is expensive. Due to the rapid developments in CCTV (Closed Circuit Television) surveillance system in recent years and video processing techniques, there is a big trend to replace conventional fire detection techniques with computer vision-based systems. This new technology can provide more reliable information and can be more cost-effective. The main objective of fire detection systems is high detection accuracy, low error rate and reasonable time detect. The video fire detection technology uses CCD cameras to capture images of the observed scene, which provides abundant and intuitive information for fire detection using image processing algorithms.
This paper presents an efficient fire detection system which detects fire areas by analyzing the videos that are acquired by surveillance cameras in urban out-door environment, especially in storage of Yazd Gas Company. Proposed method uses color, spatial and temporal information that makes a good distinction between the fire and objects which are similar to the fire.  The purpose is achieved using multi- filter. The first filter separates red color area as a primary fire candidate. The second filter operates based on the difference between fire candidate areas in the sequence of frames. In the last filter, variation of red channel in the candidate area is computed and compared with the threshold which is updating continuously. In the experiments, the performance of these filters are evaluated separately. The proposed final system is a combination of all filters. Experimental results show that the precision of the final proposed system in urban out-door environment is 100%, and our technique achieves the average detection rate of 87.35% which outperforms other base methods.