We propose a novel deep framework, TraCount, for highly overlapping vehicle counting in congested traffic scenes. TraCount uses multiple fully convolutional(FC) sub-networks to predict the density map for a given stat...
详细信息
ISBN:
(纸本)9781450347532
We propose a novel deep framework, TraCount, for highly overlapping vehicle counting in congested traffic scenes. TraCount uses multiple fully convolutional(FC) sub-networks to predict the density map for a given static image of a traffic scene. the different FC sub-networks provide a range in size of receptive fields that enable us to count vehicles whose perspective effect varies significantly in a scene due to the large visual field of surveillance cameras. the predictions of different FC sub-networks are fused by weighted averaging to obtain a final density map. We show that TraCount outperforms the state of the art methods on the challenging TRANCOS dataset that has a total of 46796 vehicles annotated across 1244 images.
Current denoising techniques use the classical orthonormal wavelets for decomposition of an image corrupted with additive white Gaussian noise, upon which various thresholding strategies are built. the use of availabl...
详细信息
ISBN:
(纸本)9781424442195
Current denoising techniques use the classical orthonormal wavelets for decomposition of an image corrupted with additive white Gaussian noise, upon which various thresholding strategies are built. the use of available biorthogonal wavelets in image denoising is less common because of their poor performance. hi this paper, we present a method to design image-matched biorthogonal wavelet bases and report on their potential for denoising. We have conducted experiments on various image datasets namely Natural, Satellite and Medical withthe designed wavelets using two existing thresholding strategies. Test results front comparing the performance of matched and fixed biorthogonal wavelets show an average improvement of 35% in MSE for low SNR values (0 to 18db) in every dataset. this improvement was also seen in the PSNR and visual comparisons. this points to the importance of matching when using wavelet-based denoising.
image based methods are a new approach for solving problems in mobile robotics. Instead of building a metric (3D) model of the environment, these methods work directly in the sensor (image) space. the environment is r...
详细信息
ISBN:
(纸本)9781424442195
image based methods are a new approach for solving problems in mobile robotics. Instead of building a metric (3D) model of the environment, these methods work directly in the sensor (image) space. the environment is represented as a topological graph in which each node contains an image taken at some pose in the workspace, and edges connect poses between which a simple path exists. this type of representation is highly scalable and is also well suited to handle the data association problems that effect metric model based methods. In this paper we present an efficient, adaptive method for qualitative localization using content based image retrieval techniques. In addition, we demonstrate an algorithm which can convert this topological graph into a metric model of the environment by incorporating information about loop closures.
the different tissues namely gray matter (GM) white matter (WM), and cerebrospinal fluid (CSF) are spread over the entire brain. It is difficult to demarcate them individually when a brain image is considered. the bou...
详细信息
ISBN:
(纸本)9781479915880
the different tissues namely gray matter (GM) white matter (WM), and cerebrospinal fluid (CSF) are spread over the entire brain. It is difficult to demarcate them individually when a brain image is considered. the boundaries are not well defined. Modified fuzzy C means (MFCM) and level sets segmentation based methodology is proposed in this paper for automated brain MRI image segmentation into WM, GM and CSF. the initial segmentation is done by MFCM approach and the results thus obtained are input to the level set methodology. We have tested the methodology on 100 different brain MRI images. the results are compared by using individual MFCM and level set segmentation methods. We took the opinion of 10 expert radiologists to corroborate our results. the results are validated by radiologists as 'Accurate', 'Satisfactory', 'Adequate' and 'Not acceptable'. the results obtained using only level set are 'not acceptable'. Most of the results obtained using MFCM are 'Adequate'. the results obtained using combined method are 'Satisfactory'. Hence, the results obtained using combined MFCM and level sets based segmentation are considered better than using individual MFCM and level set segmentation methods. the manual intervention is avoided in the combined approach. the time required to segment using combined approach is also less compared to level set method. the segmentation using proposed methodology is helpful for radiologists in hospitals for brain MRI image analysis.
Motion detection using a stationary camera can be done by estimating the static scene (background). In that purpose, we propose a new method based on a simple recursive non linear operator, the Sigma-Delta filter. Use...
详细信息
Motion detection using a stationary camera can be done by estimating the static scene (background). In that purpose, we propose a new method based on a simple recursive non linear operator, the Sigma-Delta filter. Used along with a spatiotemporal regularization algorithm, it allows robust, computationally efficient and accurate motion detection. To deal with complex scenes containing a wide range of motion models with very different time constants, we propose a generalization of the basic model to multiple Sigma-Delta estimation. (c) 2006 Elsevier B.V. All rights reserved.
Radiation therapy is a standard treatment for a large variety of cancers. Radiation treatment planning involves delineating tumor, areas-at-risk, as well as normal anatomical structures. Automated techniques for detec...
详细信息
ISBN:
(纸本)9781467385640
Radiation therapy is a standard treatment for a large variety of cancers. Radiation treatment planning involves delineating tumor, areas-at-risk, as well as normal anatomical structures. Automated techniques for detection and segmentation of tumor is a challenging task because of the absence of any standard model to identify anatomical structures. As radiation therapy is delivered in multiple daily doses (called fractions) treatment fields need to be positioned so that the daily uncertainties in the target position are covered adequately. this is done by adding a margin called the Planning Target Volume (PTV) margin around the clinical target volume. the PTV margin is determined from an understanding of the daily uncertainties. In this paper, we intend to aid doctors in determining the uncertainties of daily pelvis positioning for radiotherapy by estimating the degree of change in the pelvic bone position. We propose a technique for automatically detecting regions of pelvic bone anatomy from CBCT images. We track the change in the bone positions in pelvic regions on a daily basis. A small change suggests that the uncertainty involved in the position of tumorous cell is less, indicating that the planning target margin may be reduced substantially.
Comprehensive database that contains all possible variations of handwriting is crucial for training and recognition. the primary challenge for an optical character recognizer (OCR) is that a number of interclass chara...
详细信息
ISBN:
(纸本)9781467385640
Comprehensive database that contains all possible variations of handwriting is crucial for training and recognition. the primary challenge for an optical character recognizer (OCR) is that a number of interclass characters bear structural resemblance whereas images within a class render much dissimilarity. Acquisition of such a large database that ensures robust training of the recognizer is a painstaking task. therefore, recent research interests have been to create, from a few samples of handwriting, a comprehensive synthetic database which not only ensures naturalness, but provides much needed pattern variability. In this paper, we propose a new approach of synthetic handwritten numeral generation for Odia language using interclass deformation. We experimentally evaluate the generated databases using the state-of-the-art recognition systems. the recognition results are compared on two benchmark databases (ISI Kolkata and IIT Bhubaneswar Odia numeral) as well as two newly created synthetic databases. the Odia numeral database sizes are increased by 20-fold each using our proposed approach. the introduction of nonlinear pattern variance because of interclass deformation is proved to pose better challenge to conventional recognizers. We also experimented on a mixture of original and synthetic database for training the OCR to achieve robustness and higher accuracy.
We propose a novel physically based method to simulate explosions and other compressible fluid phenomena. the method uses compressible Navier Stokes equations for modeling the explosion with a Semi-Lagrangian integrat...
详细信息
ISBN:
(纸本)9781424442195
We propose a novel physically based method to simulate explosions and other compressible fluid phenomena. the method uses compressible Navier Stokes equations for modeling the explosion with a Semi-Lagrangian integration method. the proposed integration method addresses the issues of stability and larger timesteps. this is achieved by modifying the Semi-Lagrangian method to reduce dissipation and increase accuracy, using improved interpolation and an error correction method. the proposed method allows the rendering of related phenomena like a fireball, dust and smoke clouds, and the simulation of solid interaction like rigid fracture and rigid body simulation. Our method is flexible enough to afford substantial artistic control over the behavior of the explosion.
Cognitive load is defined as the mental workload imparted on brain while doing a task. the amount of cognitive load experienced depends on individual's ability of perception, assimilation and response to a task. R...
详细信息
ISBN:
(纸本)9781467385640
Cognitive load is defined as the mental workload imparted on brain while doing a task. the amount of cognitive load experienced depends on individual's ability of perception, assimilation and response to a task. Real-time measurement of the level of cognitive load using low cost Electroencephalogram (EEG) signal enables understanding of personal cognitive skills. In this paper, we propose a methodology of selecting a reference task whose bio-markers closely match with a given task while probing different cognitive abilities. the benefit of this approach is to have a limited set of training models for the reference tasks related to various cognitive categories and use the same for a variety of unknown tasks. Experiment is performed for two levels of cognitive load withthree different tasks namely Stroop color task, logical reasoning task and usage of on-screen keyboards. the training models of the reference tasks, selected by cluster analysis of low and high cognitive levels are used to evaluate an unknown task. Experimental results indicate that the Stroop is a better reference for On-Screen keyboard test compared to the Logical reasoning test. Support vector machine (SVM) and principal component analysis (PCA) followed by SVM (PCASVM) are used as the classifiers for the testing.
the increasing urban population growth leads to challenges in cities in many aspects: Urbanisation problems such as excessive environmental pollution or increasing urban traffic demand new and innovative solutions. In...
详细信息
暂无评论