The Infona portal uses cookies, i.e. strings of text saved by a browser on the user's device. The portal can access those files and use them to remember the user's data, such as their chosen settings (screen view, interface language, etc.), or their login data. By using the Infona portal the user accepts automatic saving and using this information for portal operation purposes. More information on the subject can be found in the Privacy Policy and Terms of Service. By closing this window the user confirms that they have read the information on cookie usage, and they accept the privacy policy and the way cookies are used by the portal. You can change the cookie settings in your browser.
Understanding a scene provided by very high resolution (VHR) satellite imagery has become a more and more challenging problem. In this paper, we propose a new method for scene classification based on saliency computing of patches sampling from the VHR images. Sparse principal component analysis (sPCA) is then adopted to select the corresponding informative salient patches for image scene representation...
In computer vision applications such as person re-identification the optimization of rank list is an important issue. In order to address this issue, a multi-feature fusion based re-ranking technique is proposed. In most of the conventional methods, a long feature vector is formulated from a single modality. Whereas, in the proposed approach, multiple features from the image are extracted and combined...
Many emerging motion-related applications, such as virtual reality, decision making, and health monitoring, demand reliability and quick response upon input changes. Motion capture has been a well-researched topic in the past decades with applications in many industries. The ability to capture motion goes hand in hand with real-time capability in a system. This paper gives an overview on real-time...
This paper proposes image super-resolution techniques with multi-channel convolutional neural networks (CNN). In the proposed method, output pixels are classified into four groups depending on their positions. Those groups are generated from separate channels of the CNN. Finally, they are synthesized into a 2−2 magnified image. This architecture can enlarge images directly without bicubic interpolation...
Even though there has been enormous research in facial analysis and more sophisticated algorithm, face recognition fails drastically in real time when the facial images are occluded. This paper explains the algorithm and technical concepts behind the high accurate face recognition systems for a 2D frontal images with occlusion for a business requirments. Face recognition is implemented using Convolutional...
Three Dimensional reconstruction of an object is one of the current research topics. This paper discusses the generation of a model in Three-Dimensional space by taking views of the object from different angles. Image acquisition is done using a single digital camera and turn table. Camera calibration is done using the images of checkerboard pattern. Camera calibration gives the intrinsic and extrinsic...
In this work, we develop a video surveillance system to detect the disappearance of the objects selected by an operator. Proposed method is optimized to minimize the effects of shadows, partial occlusions and changes in illumination to minimize false alarms. The system is trained by extracting local features from the object to produce an alarm in the case of the disappearance.
The more digital image production increase, the more effectiveness and productivity of image recognition mechanism and comparison techniques become important. In this study, comparative experiments are made for distribution distance measurement function that is explained in the scope by using skeleton presentation which is obtained from 2D and 3D images of objects. Separate experiments are made in...
Today, for two or three dimensional modelling of objects, common algorithms such as spline curves and triangulation methods are being used. However, while these methods are able to well express the objects visually, since they approach the objects as a composition of a large number of small pieces, they are not able to cognisably put forth the mathematical expression of the object as a whole. The...
In this paper, a new shaped marker is designed for augmented reality applications that has 360 degree viewing angle about its shaft axis. The main advantage of the designed marker is that it is very simple to extract marker area from images with basic image processing methods. And decoding of the marker codes is very simple with basic mathematical functions. Experimental results showed that, the designed...
Learning from imbalanced data sets is an important problem frequently encountered in the application of classification problems. Instances of this type of problem is usually labeled with the label of class majority and minority class instances will be ignored. In this study, an ensemble based method is proposed for problems of imbalanced data set. The results obtained were compared to alternative...
Public safety has became an important issue in recent years. Developing smart systems to detect abnormal crowd behavior is crucial to intervene the situation as soon as possible. In this work, we propose a novel algorithm based on Coarse to Fine Optical Flow and Influence Map.
In work, a real-time speed limit warning assistant system for drivers is presented. The proposed approach is able to produce optical and acoustic warnings by detecting speed limit signs simply processing captured images via a camera. The system developed can process by up to 7 frame per second on Raspberry Pi 2 single board computer.
In this paper, using a general convolutional neural network (CNN) model, which was developed for object recognition, a successful system has been introduced for the person re-identification problem. To use this CNN model for the person re-identification problem properly, it is individually fine-tuned using different body parts of person images. For feature extraction, we used the seventh layer of...
In this study, the automated matching of 2.5 m resolution Göktürk-2 panchromatic stereo images has been addressed. From an operational perspective, it seems unlikely to produce the epipolar images from Göktürk-2 stereo datasets at a sub-pixel level due to several reasons. Therefore, SIFT-flow method that does not require any user input and that has ability to perform matching through the stereo data...
Today, trying to understand what kind of behaviour the crowd shows by studying the data from surveillance systems is an important topic for researchers of computer vision. The aim of this study make the motion data that is at pixel level and that is obtained by optical flow method a more meaningful data set with the particle advection method. In other words, the aim is to monitor the motion data by...
Deep convolutional neural networks is a recently developed method that yields very successful results in image classification. Deep neural networks, which have a high number of parameters, require a large amount of data to avoid overfitting during training. For applications in which the available data is not adequate to train a deep neural network from the scratch, deep neural networks trained for...
In search and rescue operations, it is severely crucial to perform a quick scan in the disaster area, localize the wounded people and send aid. In this study, approaches for mapping an unknown environment using image stitching technique was investigated. The image map was employed for detecting humans using computer vision techniques. For indoor environments where GPS is not available, research showed...
We propose an unsupervised method for abnormal crowd activity detection in surveillance systems. Proposed solution is using MPEG-7 Motion Activity descriptors and Particle Filter algorithm for classification. The experiments were performed on UMN dataset sequences. The detection results are comparable to results obtained by supervised methods.
We propose a systematic frame work for the automatic detection of multiple human actions within the same frame in realistic and diverse video settings. One of the major challenges is the process of recognizing and understanding of human actions from videos with large variations resulting from camera motions, changes in human appearance, pose changes, scale changes and back ground clutter etc. In this...
Set the date range to filter the displayed results. You can set a starting date, ending date or both. You can enter the dates manually or choose them from the calendar.