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.
In this paper, we present an effective method to recognize human actions from sequences of depth maps, which are captured by a consume depth sensor. In our approach, we first project each frame of a depth sequence onto three orthogonal planes and generate the depth motion sequence (DMS) between two consecutive frames from the three projected views. Then we propose a spatio-temporal cuboid pyramid...
It is very important to realize the safer marine transportation system. This paper describes a novel method to detect and tracking ships using a machine vision to reduce marine accidents. The proposed method has two stages; detection and tracking. In the detection stage, edges on a sea image are calculated and bounding rectangles that have the many pixels with the strong edge are selected by morphological...
A simple and robust gesture recognition system is proposed for better human-computer interaction using Microsoft's Kinect sensor. The Kinect is employed to construct skeletons for a subject in the 3D space using twenty body joint coordinates. From this skeletal information, ten joints are required and six triangles have been constructed along with six respective centroids. The feature space corresponds...
In order to improve the ability of arithmetic of school children, we develop a real-time magic square solver based on NVIDIA Jetson TK1, which recognizes user's puzzle images and then shows several hints and an answer of the puzzle in real-time. We also confirm the effectiveness of image recognition and our puzzle solver by experiments.
Recently, support vector ranking has been adopted to address the challenging person re-identification problem. However, the ranking model based on ordinary global features cannot represent the significant variation of pose and viewpoint across camera views. Thus, a novel ranking method which fuses the dense invariant features is proposed in this paper to model the variation of images across camera...
In this paper, a deep convolutional neural network based approach to the problem of automatically recognizing jersey numbers from soccer videos is presented. Two different jersey number vector encoding schemes are presented and compared to each other. The first treats every number as a separate class, while the second one treats each digit as a class. Additionally, the semi-automatic process for the...
License plate recognition (LPR) is investigated as a promising automation technique for transportation system, for instance, automatic billing system for electronic parking lots, snapshot system of traffic violation, etc. Support Vector Machine (SVM) is a typical method for pattern recognition, and it works efficiently for traditional vehicle plate recognition systems, which process the vehicle images...
Particulate pollution has become increasingly critical and threatening for human health. Although a number of approaches have been attempted for particulate pollution monitoring, these approaches are either expensive, unscalable, or requiring deployment of yet-another sensing infrastructure. In this study, by combining the advanced image dehazing and support vector machine techniques, we propose a...
This paper presents a novel method for eye tracking and blink detection in the video frames obtained from low resolution consumer grade web cameras. It uses a method involving Haar based cascade classifier for eye tracking and a combination of HOG features with SVM classifier for eye blink detection. The presented method is non intrusive and hence provides a comfortable user interaction. The eye tracking...
This paper presents an accurate method of drowsiness detection for the images obtained using low resolution consumer grade web cameras under normal lighting conditions. The drowsiness detection method uses Haar based cascade classifier for eye tracking and combination of Histogram of oriented gradient (HOG) features combined with Support Vector Machine (SVM) classifier for blink detection. Once the...
In the technology of Advanced Safety Vehicle(ASV), it is important elements to detect pedestrians. There is a technique which used image features and a classifier for pedestrian detection. However, pedestrians have pose variations, such as a rotation. That's why we cannot always get the same features. In order to solve this problem, we can use a way which used the rotation-invariant HOG, but this...
In this paper we introduce an assistive device dedicated to visual impaired / blind people completely integrated on a regular smartphone. The framework is designed to detect and localize static and dynamic obstacle during user navigation. We start by selecting a reduced and relevant set of FAST interest points based on a regular grid and Harris-Laplacian operator. Then, we construct a global image...
The aim of this work is to propose a fusion procedure based on lidar and camera to solve the pedestrian detection problem in autonomous driving. Current pedestrian detection algorithms have focused on improving the discriminability of 2D features that capture the pedestrian appearance, and on using various classifier architectures. However, less focus on exploiting the 3D structure of object has limited...
One key issue for people re-identification is to find good features or representation to bridge the gaps among different appearances of the same people, which is introduced by large variances in view point, illumination and non-rigid deformation. In this paper, we create a deep convolutional neural network (deep CNN) to solve this problem and integrate feature learning and re-identification into one...
Although a great success has been achieved on action detection tasks by using "bag of features" architecture as video representations, action detection with web camera still remains a challenge. Most of these algorithms can extract features either sparsely at interest points or densely on regular grids, usually, sampling densely can get better results than sampling sparsely using the local...
This paper presents an empirical study on mining comparative sentences for Vietnamese language. Given a set of evaluative Vietnamese documents, the goal of comparative sentence mining consists of (1) identifying comparative sentences in the documents, and (2) recognition of relations in identified comparative sentences. A relation describes a comparison of two entities or two sets of entities in some...
Human action recognition in wild scene is discussed and a novel approach of dense trajectory selection is addressed in this paper, in order to deal with side effects ascribed to clutter, unsteady background interference, camera motion and random noise in the video. First, multi-scale temporal pyramid is constructed from original frames in the video. By employing dense sampling, candidate initial points...
In this paper, a deep-sea high-definition camera system with the function of sea creatures' recognition is designed. The system is improved based on the traditional deep-sea camera system; it can not only have a real-time monitoring of water environment but also improves the speed of data transmission and video image resolution. Image processing technique is also applied in this system. We can identify...
Rotary-wing unmanned aerial vehicles (UAV) are being widely used in different applications due to its several features, such as mobility, lightweight, embedded processing and capability of flying in different height levels. Among the possible applications they are used in surveillance tasks, agriculture environments monitoring, power lines inspections and diseases detection in crops. The images captured...
It was proved that the fusion of information from multi-modality images increases the accuracy of pedestrian recognition systems. One of the best approache so far is to concatenate the features from multi-modality images into a large feature vector, but it requires strong camera calibration settings and non-discriminative modalities could lead to missclassification of some particular images. We present...
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.