Search results for: saliency detection

Authors: Junling Li, Fang Meng, Yichun Zhang

Abstract:

Image saliency detection has been long studied, while several challenging problems are still unsolved, such as detecting saliency inaccurately in complex scenes or suppressing salient objects in the image borders. In this paper, we propose a new saliency detection algorithm in order to solving these problems. We represent the image as a graph with superixels as nodes. By considering appearance similarity between the boundary and the background, the proposed method chooses non-saliency boundary nodes as background priors to construct the background probability model. The probability that each node belongs to the model is computed, which measures its similarity with backgrounds. Thus we can calculate saliency by the transformed probability as a metric. We compare our algorithm with ten-state-of-the-art salient detection methods on the public database. Experimental results show that our simple and effective approach can attack those challenging problems that had been baffling in image saliency detection.

Keywords: visual saliency, background probability, boundary knowledge, background priors

Procedia PDF Downloads 406

Authors: Natalia Efremova, Sergey Tarasenko

Abstract:

We propose a novel neural network architecture for visual saliency detections, which utilizes neuro physiologically plausible mechanisms for extraction of salient regions. The model has been significantly inspired by recent findings from neuro physiology and aimed to simulate the bottom-up processes of human selective attention. Two types of features were analyzed: color and direction of maximum variance. The mechanism we employ for processing those features is PCA, implemented by means of normalized Hebbian learning and the waves of spikes. To evaluate performance of our model we have conducted psychological experiment. Comparison of simulation results with those of experiment indicates good performance of our model.

Keywords: neural network models, visual saliency detection, normalized Hebbian learning, Oja's rule, psychological experiment

Procedia PDF Downloads 326

Authors: Tanveer Hussain, Khan Muhammad, Amin Ullah, Mi Young Lee, Sung Wook Baik

Abstract:

The Big Data generated from distributed vision sensors installed on large scale in smart cities create hurdles in its efficient and beneficial exploration for browsing, retrieval, and indexing. This paper presents a three-folded framework for effective video summarization of such data and provide a compact and representative format of Big Video Data. In the first fold, the paper acquires input video data from the installed cameras and collect clues such as type and count of objects and clarity of the view from a chunk of pre-defined number of frames of each view. The decision of representative view selection for a particular interval is based on fuzzy inference system, acquiring a precise and human resembling decision, reinforced by the known clues as a part of the second fold. In the third fold, the paper forwards the selected view frames to the summary generation mechanism that is supported by a saliency-aware convolution neural network (CNN) model. The new trend of fuzzy rules for view selection followed by CNN architecture for saliency computation makes the multi-view video summarization (MVS) framework a suitable candidate for real-world practice in smart cities.

Keywords: big video data analysis, fuzzy logic, multi-view video summarization, saliency detection

Procedia PDF Downloads 169

Authors: Ahmad Absetan, Mahdi Nooshyar

Abstract:

The importance map is an image-based measure and is a core part of the resizing algorithm. Importance measures include image gradients, saliency and entropy, as well as high level cues such as face detectors, motion detectors and more. In this work we proposed a new method to calculate the importance map, the importance map is generated automatically using a novel combination of image edge density and Harel saliency measurement. Experiments of different type images demonstrate that our method effectively detects prominent areas can be used in image resizing applications to aware important areas while preserving image quality.

Keywords: content-aware image resizing, visual saliency, edge density, image warping

Procedia PDF Downloads 557

Authors: Chin-Chen Chang, I-Ta Lee, Tsung-Ta Ke, Wen-Kai Tai

Abstract:

Common methods for resizing image size include scaling and cropping. However, these two approaches have some quality problems for reduced images. In this paper, we propose an image resizing algorithm by separating the main objects and the background. First, we extract two feature maps, namely, an enhanced visual saliency map and an improved gradient map from an input image. After that, we integrate these two feature maps to an importance map. Finally, we generate the target image using the importance map. The proposed approach can obtain desired results for a wide range of images.

Keywords: energy map, visual saliency, gradient map, seam carving

Procedia PDF Downloads 463

Authors: William Huang

Abstract:

Retinal prostheses are currently limited to low resolution grayscale images that lack color and spatial information. This study develops a novel real-time image optimization framework and tools to encode maximum information to the prostheses which are constrained by the number of electrodes. One key idea is to localize main objects in images while reducing unnecessary background noise through region-contrast saliency maps. A novel color depth mapping technique was developed through MiniBatchKmeans clustering and color space selection. The resulting image was downsampled using bicubic interpolation to reduce image size while preserving color quality. In comparison to current schemes, the proposed framework demonstrated better visual quality in tested images. The use of the region-contrast saliency map showed improvements in efficacy up to 30%. Finally, the computational speed of this algorithm is less than 380 ms on tested cases, making real-time retinal prostheses feasible.

Keywords: retinal implants, virtual processing unit, computer vision, saliency maps, color quantization

Procedia PDF Downloads 125

Authors: Chao Yu, Qing-Guo Wang, Dan Zhang

Abstract:

The outlier detection problem for dynamic systems is formulated as a matrix decomposition problem with low-rank, sparse matrices and further recast as a semidefinite programming (SDP) problem. A fast algorithm is presented to solve the resulting problem while keeping the solution matrix structure and it can greatly reduce the computational cost over the standard interior-point method. The computational burden is further reduced by proper construction of subsets of the raw data without violating low rank property of the involved matrix. The proposed method can make exact detection of outliers in case of no or little noise in output observations. In case of significant noise, a novel approach based on under-sampling with averaging is developed to denoise while retaining the saliency of outliers and so-filtered data enables successful outlier detection with the proposed method while the existing filtering methods fail. Use of recovered “clean” data from the proposed method can give much better parameter estimation compared with that based on the raw data.

Keywords: outlier detection, system identification, matrix decomposition, low-rank matrix, sparsity, semidefinite programming, interior-point methods, denoising

Procedia PDF Downloads 290

Authors: Yili Chen, Xiaokun Liang, Zhicheng Zhang, Yaoqin Xie

Abstract:

In an effort to restore visual perception in retinal diseases, an electronic retinal prosthesis with thousands of electrodes has been developed. The image processing strategies of retinal prosthesis system converts the original images from the camera to the stimulus pattern which can be interpreted by the brain. Practically, the original images are with more high resolution (256x256) than that of the stimulus pattern (such as 25x25), which causes a technical image processing challenge to do large-capacity image information reduction. In this paper, we focus on developing an efficient image processing stimulus pattern extraction algorithm by using a single cue saliency map for extracting salient objects in the image with an optimal trimming threshold. Experimental results showed that the proposed stimulus pattern extraction algorithm performs quite well for different scenes in terms of the stimulus pattern. In the algorithm performance experiment, our proposed SCSPE algorithm have almost five times of the score compared with Boyle’s algorithm. Through experiment s we suggested that when there are salient objects in the scene (such as the blind meet people or talking with people), the trimming threshold should be set around 0.4max, in other situations, the trimming threshold values can be set between 0.2max-0.4max to give the satisfied stimulus pattern.

Keywords: retinal prosthesis, image processing, region of interest, saliency map, trimming threshold selection

Procedia PDF Downloads 224

Authors: Xiaoying Guo, Xiangyun Wang, Chunhua Jia

Abstract:

Web page has been one of the most important way of knowing the world. Humans catch a lot of information from it everyday. Thus, understanding where human looks when they surfing the web pages is rather important. In normal scenes, the down-top features and top-down tasks significantly affect humans’ eye movement. In this paper, we investigated if the conventional visual salience algorithm can properly predict humans’ visual attractive region when they viewing the web pages. First, we obtained the eye movement data when the participants viewing the web pages using an eye tracker. By the analysis of eye movement data, we studied the influence of visual saliency and thinking way on eye-movement pattern. The analysis result showed that thinking way affect human’ eye-movement pattern much more than visual saliency. Second, we compared the results of web page visual salience region extracted by Itti model and Spectral Residual (SR) model. The results showed that Spectral Residual (SR) model performs superior than Itti model by comparison with the heat map from eye movements. Considering the influence of mind habit on humans’ visual region of interest, we introduced one of the most important cue in mind habit-fixation position to improved the SR model. The result showed that the improved SR model can better predict the human visual region of interest in web pages.

Keywords: web page salience region, eye-tracker, spectral residual, visual salience

Procedia PDF Downloads 257

Authors: Dingyang Hu, Dan Liu

Abstract:

DNN (Deep Neural Network) deep learning models are widely used in classification, prediction, and other task scenarios. To address the difficulties of generic adversarial perturbation generation for deep learning models under black-box conditions, a generic adversarial ingestion generation method based on a saliency map (CJsp) is proposed to obtain salient image regions by counting the factors that influence the input features of an image on the output results. This method can be understood as a saliency map attack algorithm to obtain false classification results by reducing the weights of salient feature points. Experiments also demonstrate that this method can obtain a high success rate of migration attacks and is a batch adversarial sample generation method.

Keywords: adversarial sample, gradient, probability, black box

Procedia PDF Downloads 70

Authors: Christoph Opperer

Abstract:

Space syntax is a theory and method for analyzing the spatial layout of buildings and urban environments to understand how they can influence patterns of human movement, social interaction, and behavior. While direct visibility is a key factor in space syntax research, important visual information such as light, color, texture, etc., are typically not considered, even though psychological studies have shown a strong correlation to the human perceptual experience within physical space – with light and color, for example, playing a crucial role in shaping the perception of spaciousness. Furthermore, these surface properties are often the visual features that are most salient and responsible for drawing attention to certain elements within the environment. This paper explores the potential of integrating these factors into general space syntax methods and visibility-based analysis of space, particularly for architectural spatial layouts. To this end, we use a combination of geometric (isovist) and topological (visibility graph) approaches together with image-based methods, allowing a comprehensive exploration of the relationship between spatial geometry, visual aesthetics, and human experience. Custom-coded ray-tracing techniques are employed to generate spherical panorama images, encoding three-dimensional spatial data in the form of two-dimensional images. These images are then processed through computer vision algorithms to generate saliency-maps, which serve as a visual representation of areas most likely to attract human attention based on their visual properties. The maps are subsequently used to weight the vertices of isovists and the visibility graph, placing greater emphasis on areas with high saliency. Compared to traditional methods, our weighted visibility analysis introduces an additional layer of information density by assigning different weights or importance levels to various aspects within the field of view. This extends general space syntax measures to provide a more nuanced understanding of visibility patterns that better reflect the dynamics of human attention and perception. Furthermore, by drawing parallels to traditional isovist and VGA analysis, our weighted approach emphasizes a crucial distinction, which has been pointed out by Ervin and Steinitz: the difference between what is possible to see and what is likely to be seen. Therefore, this paper emphasizes the importance of including surface properties in visibility-based analysis to gain deeper insights into how people interact with their surroundings and to establish a stronger connection with human attention and perception.

Keywords: space syntax, visibility analysis, isovist, visibility graph, visual features, human perception, saliency detection, raytracing, spherical images

Procedia PDF Downloads 50

Authors: Chee Sun Won

Abstract:

This paper presents the size reduction of images for machine-to-machine communications. Here, the salient image regions to be preserved include the image patches of the key-points such as corners and blobs. Based on a saliency image map from the key-points and their image patches, an axis-aligned grid-size optimization is proposed for the reduction of image size. To increase the size-reduction efficiency the aspect ratio constraint is relaxed in the constraint optimization framework. The proposed method yields higher matching accuracy after the size reduction than the conventional content-aware image size-reduction methods.

Keywords: image compression, image matching, key-point detection and description, machine-to-machine communication

Procedia PDF Downloads 397

Authors: Yili Chen, Jixiang Fu, Zhihua Liu, Zhicheng Zhang, Rongmao Li, Nan Fu, Yaoqin Xie

Abstract:

Retinal prothesis is designed to help the blind to get some sight.It is made up of internal part and external part.In external part ,there is made up of camera, image processing, and RF transmitter.In internal part, there is RF receiver, implant chip,micro-electrode.The image got from the camera should be processed by suitable stragies to corresponds to stimulus the electrode.Nowadays, the number of the micro-electrode is hundreds and we don’t know the mechanism how the elctrode stimulus the optic nerve, an easy way to the hypothesis is that the pixel in the image is correspondence to the electrode.So it is a question how to get the important information of the image captured from the picture.There are many strategies to experimented to get the most important information as soon as possible, due to the real time system.ROI is a useful algorithem to extract the region of the interest.Our paper will explain the details of the orinciples and functions of the ROI.And based on this, we simplified the ROI algrithem,and used it in outside image prcessing DSP system of the retinal prothesis.Results show that our image processing stratiges is suitable for real-time retinal prothesis and can cut redundant information and help useful information to express in the low-size image.

Keywords: image processing, region of interest, saliency map, low-size image, useful information express, cut redundant information in image

Procedia PDF Downloads 257

Authors: Jae-Jeong Kim, Ki-Ro Kim, Hyoung-Kyu Song

Abstract:

In this paper, we propose an efficient signal detector that switches M parameter of QRD-M detection scheme is proposed for MIMO-OFDM system. The proposed detection scheme calculates the threshold by 1-norm condition number and then switches M parameter of QRD-M detection scheme according to channel information. If channel condition is bad, the parameter M is set to high value to increase the accuracy of detection. If channel condition is good, the parameter M is set to low value to reduce complexity of detection. Therefore, the proposed detection scheme has better trade off between BER performance and complexity than the conventional detection scheme. The simulation result shows that the complexity of proposed detection scheme is lower than QRD-M detection scheme with similar BER performance.

Keywords: MIMO-OFDM, QRD-M, channel condition, BER

Procedia PDF Downloads 344

Authors: Jae-Hyun Ro, Yong-Jun Kim, Chang-Bin Ha, Hyoung-Kyu Song

Abstract:

In multiple input multiple output-orthogonal frequency division multiplexing (MIMO-OFDM) systems, many detection schemes have been developed to improve the error performance and to reduce the complexity. Maximum likelihood (ML) detection has optimal error performance but it has very high complexity. Thus, this paper proposes reduced complexity of ML detection combined with decision feedback equalizer (DFE). The error performance of the proposed detection scheme is higher than the conventional DFE. But the complexity of the proposed scheme is lower than the conventional ML detection.

Keywords: detection, DFE, MIMO-OFDM, ML

Procedia PDF Downloads 591

Authors: Wei Li, Tuo Yang

Abstract:

In order to satisfy the real-time and accurate requirements of cigarette smoke detection in complex scenes, a cigarette smoke detection technology based on the combination of deep learning and color features was proposed. Firstly, based on the color features of cigarette smoke, the suspicious cigarette smoke area in the image is extracted. Secondly, combined with the efficiency of cigarette smoke detection and the problem of network overfitting, a network model for cigarette smoke detection was designed according to YOLOV3 algorithm to reduce the false detection rate. The experimental results show that the method is feasible and effective, and the accuracy of cigarette smoke detection is up to 99.13%, which satisfies the requirements of real-time cigarette smoke detection in complex scenes.

Keywords: deep learning, computer vision, cigarette smoke detection, YOLOV3, color feature extraction

Procedia PDF Downloads 65

Authors: H. Benmoussa, A. A. El Kalam, A. Ait Ouahman

Abstract:

The design and implementation of intrusion detection systems (IDS) remain an important area of research in the security of information systems. Despite the importance and reputation of the current intrusion detection systems, their efficiency and effectiveness remain limited as they should include active defense approach to allow anticipating and predicting intrusions before their occurrence. Consequently, they must be readapted. For this purpose we suggest a new generation of distributed intrusion detection system based on preventive detection approach and using intelligent and mobile agents. Our architecture benefits from mobile agent features and addresses some of the issues with centralized and hierarchical models. Also, it presents advantages in terms of increasing scalability and flexibility.

Keywords: Intrusion Detection System (IDS), preventive detection, mobile agents, distributed architecture

Procedia PDF Downloads 557

Authors: Omair Ghori, Anton Stadler, Stefan Wilk, Wolfgang Effelsberg

Abstract:

Fire-related incidents account for extensive loss of life and material damage. Quick and reliable detection of occurring fires has high real world implications. Whereas a major research focus lies on the detection of outdoor fires, indoor camera-based fire detection is still an open issue. Cameras in combination with computer vision helps to detect flames and smoke more quickly than conventional fire detectors. In this work, we present a computer vision-based smoke detection algorithm based on contrast changes and a multi-step classification. This work accelerates computer vision-based fire detection considerably in comparison with classical indoor-fire detection.

Keywords: contrast analysis, early fire detection, video smoke detection, video surveillance

Procedia PDF Downloads 417

Authors: Rashid Mahmood

Abstract:

The network as a service (NaaS) usage has been well-known from the last few years in the many applications, like mission critical applications. In the NaaS, prevention method is not adequate as the security concerned, so the detection method should be added to the security issues in NaaS. The authentication and encryption are considered the first solution of the NaaS problem whereas now these are not sufficient as NaaS use is increasing. In this paper, we are going to present the concept of intrusion detection and then survey some of major intrusion detection techniques in NaaS and aim to compare in some important fields.

Keywords: IDS, cloud, naas, detection

Procedia PDF Downloads 299

Authors: Yousef Farhaoui

Abstract:

An IDS is a tool which is used to improve the level of security. We present in this paper different architectures of IDS. We will also discuss measures that define the effectiveness of IDS and the very recent works of standardization and homogenization of IDS. At the end, we propose a new model of IDS called BiIDS (IDS Based on the two principles of detection) for securing web servers and applications by the Intrusion Detection System (IDS).

Keywords: intrusion detection, architectures, characteristic, tools, security, web server

Procedia PDF Downloads 395

Authors: Ji-Hoon Hong, Dong-Hee Kim, Nam-Uk Kim, Tai-Myoung Chung

Abstract:

Smartphone users are increasing rapidly. Accordingly, many companies are running BYOD (Bring Your Own Device: Policies to bring private-smartphones to the company) policy to increase work efficiency. However, smartphones are always under the threat of malware, thus the company network that is connected smartphone is exposed to serious risks. Most smartphone malware detection techniques are to perform an independent detection (perform the detection of a single target application). In this paper, we analyzed a variety of intrusion detection techniques. Based on the results of analysis propose an agent using the network IDS.

Keywords: android malware detection, software-defined network, interaction environment, android malware detection, software-defined network, interaction environment

Procedia PDF Downloads 413

Authors: Medjram Sofiane, Babahenini Mohamed Chaouki, Mohamed Benali Yamina

Abstract:

Skin detection is an important task for computer vision systems. A good method for skin detection means a good and successful result of the system. The colour is a good descriptor that allows us to detect skin colour in the images, but because of lightings effects and objects that have a similar colour skin, skin detection becomes difficult. In this paper, we proposed a method using the YCbCr colour space for skin detection and lighting effects elimination, then we use the information of texture to eliminate the false regions detected by the YCbCr colour skin model.

Keywords: skin detection, YCbCr, GLCM, texture, human skin

Procedia PDF Downloads 429

Authors: Zhang Xiaodong, Tang Zixin, Liu Xin

Abstract:

In order to avoid self-collision of space manipulators during operation process, a real-time detection method is proposed in this paper. The manipulator is fitted into a cylinder enveloping surface, and then the detection algorithm of collision between cylinders is analyzed. The collision model of space manipulator self-links can be detected by using this algorithm in real-time detection during the operation process. To ensure security of the operation, a safety threshold is designed. The simulation and experiment results verify the effectiveness of the proposed algorithm for a 7-DOF space manipulator.

Keywords: space manipulator, collision detection, self-collision, the real-time collision detection

Procedia PDF Downloads 443

Authors: Norzalina Othman, Nurul Na’imy Wan, Azliza Mohd Rusli, Wan Noor Syahirah Meor Idris

Abstract:

Iris detection is a process where the position of the eyes is extracted from the face images. It is a current method used for many applications such as for security purpose and drowsiness detection. This paper proposes the use of eyes detection in controlling side mirror of motor vehicles. The eyes detection method aims to make driver easy to adjust the side mirrors automatically. The system will determine the midpoint coordinate of eyes detection on RGB (color) image and the input signal from y-coordinate will send it to controller in order to rotate the angle of side mirror on vehicle. The eye position was cropped and the coordinate of midpoint was successfully detected from the circle of iris detection using Viola Jones detection and circular Hough transform methods on RGB image. The coordinate of midpoint from the experiment are tested using controller to determine the angle of rotation on the side mirrors.

Keywords: iris detection, midpoint coordinates, RGB images, side mirror

Procedia PDF Downloads 402

Authors: Leping Chen, Daoxiang An, Xiaotao Huang

Abstract:

Automatic vehicle detection using synthetic aperture radar (SAR) image has been widely researched, as well as using optical remote sensing images. However, most researches treat the detection as an independent problem, failing to make full use of SAR data information. In circular SAR (CSAR), the two long borders of vehicle will shrink if the imaging surface is set higher than the reference one. Based on above variance, an automatic vehicle detection using CSAR image is proposed to enhance detection ability under complex environment, such as vehicles’ closely packing, which confuses the detector. The detection method uses the multiple images generated by different height plane to obtain an energy-concentrated image for detecting and then uses the maximally stable extremal regions method (MSER) to detect vehicles. A result of vehicles’ detection is given to verify the effectiveness and correctness of proposed method.

Keywords: circular SAR, vehicle detection, automatic, imaging

Procedia PDF Downloads 346

Authors: Bouchemha Amel, Chachoui Takieddine, H. Maalem

Abstract:

Automatic detection of targets in a modern communication system RADAR is based primarily on the concept of adaptive CFAR detector. To have an effective detection, we must minimize the influence of disturbances due to the clutter. The detection algorithm adapts the CFAR detection threshold which is proportional to the average power of the clutter, maintaining a constant probability of false alarm. In this article, we analyze the performance of two variants of adaptive algorithms CA-CFAR and OS-CFAR and we compare the thresholds of these detectors in the marine environment (no-Gaussian) with a Weibull distribution.

Keywords: CFAR, threshold, clutter, distribution, Weibull, detection

Procedia PDF Downloads 561

Authors: Rashid Mahmood, Muhammad Junaid Sarwar

Abstract:

Mobile ad hoc networks (MANETs) use has been well-known from the last few years in the many applications, like mission critical applications. In the (MANETS) prevention method is not adequate as the security concerned, so the detection method should be added to the security issues in (MANETs). The authentication and encryption is considered the first solution of the MANETs problem where as now these are not sufficient as MANET use is increasing. In this paper we are going to present the concept of intrusion detection and then survey some of major intrusion detection techniques in MANET and aim to comparing in some important fields.

Keywords: MANET, IDS, intrusions, signature, detection, prevention

Procedia PDF Downloads 353

Authors: Sulaiman Al amro, Ali Alkhalifah

Abstract:

The growing number of computer viruses and the detection of zero day malware have been the concern for security researchers for a large period of time. Existing antivirus products (AVs) rely on detecting virus signatures which do not provide a full solution to the problems associated with these viruses. The use of logic formulae to model the behaviour of viruses is one of the most encouraging recent developments in virus research, which provides alternatives to classic virus detection methods. In this paper, we proposed a comparative study about different virus detection techniques. This paper provides the advantages and drawbacks of different detection techniques. Different techniques will be used in this paper to provide a discussion about what technique is more effective to detect computer viruses.

Keywords: computer viruses, virus detection, signature-based, behaviour-based, heuristic-based

Procedia PDF Downloads 453

Authors: Kaewmart Pongakkasira

Abstract:

This study investigated how different levels of pixelation affect face detection in natural scenes. Eye movements and reaction times, while observers searched for faces in natural scenes rendered in different ranges of pixels, were recorded. Detection performance for coarse visual detail at lower pixel size (3 x 3) was better than with very blurred detail carried by higher pixel size (9 x 9). The result is consistent with the notion that face detection relies on gross detail information of face-shape template, containing crude shape structure and features. In contrast, detection was impaired when face shape and features are obscured. However, it was considered that the degradation of scenic information might also contribute to the effect. In the next experiment, a more direct measurement of the effect of pixelation on face detection, only the embedded face photographs, but not the scene background, will be filtered.

Keywords: eye movements, face detection, face-shape information, pixelation

Procedia PDF Downloads 297

Authors: M. Ranjeeth, S. Anuradha

Abstract:

Spectrum sensing is the main feature of cognitive radio technology. Spectrum sensing gives an idea of detecting the presence of the primary users in a licensed spectrum. In this paper we compare the theoretical results of detection probability of different fading environments like Rayleigh, Rician, Nakagami-m fading channels with the simulation results using energy detection based spectrum sensing. The numerical results are plotted as P_f Vs P_d for different SNR values, fading parameters. It is observed that Nakagami fading channel performance is better than other fading channels by using energy detection in spectrum sensing. A MATLAB simulation test bench has been implemented to know the performance of energy detection in different fading channel environment.

Keywords: spectrum sensing, energy detection, fading channels, probability of detection, probability of false alarm

Procedia PDF Downloads 513