Compact representation of visual content has emerged as an important topic in the context of large scale image/video retrieval. The recently proposed Vector of Locally Aggregated Descriptors (VLAD) has shown to outper...
详细信息
Compact representation of visual content has emerged as an important topic in the context of large scale image/video retrieval. The recently proposed Vector of Locally Aggregated Descriptors (VLAD) has shown to outperform other existing techniques for retrieval. In this paper, we propose two spatio-temporal features for constructing VLAD vectors for videos in the context of large scale video retrieval. Given a particular query video, our aim is to retrieve similar videos from the database. Experiments are conducted on UCF50 and HMDB51 datasets, which pose challenges in the form of camera motion, view-point variation, large intra-class variation, etc. The paper proposes the following two spatio-temporal features for constructing VLADs i) Local Histogram of Oriented Optical Flow (LHOOF), and ii) Space-Time Invariant Points (STIP). The performance of these proposed features are compared with SIFT based spatial feature. The mean average precision (MAP) indicates the better retrieval performance of the proposed spatio-temporal feature over spatial feature.
Large space with many cameras require huge storage and computational power to process these data for surveillance applications. In this paper we propose a distributed camera and processing based face detection and rec...
详细信息
Large space with many cameras require huge storage and computational power to process these data for surveillance applications. In this paper we propose a distributed camera and processing based face detection and recognition system which can generate information for finding spatiotemporal movement pattern of individuals over a large monitored space. The system is built upon Hadoop Distributed File System using map reduce programming model. A novel key generation scheme using distance based hashing technique has been used for distribution of the face matching task. Experimental results have established effectiveness of the technique.
Online handwriting recognition research has recently received significant thrust. Specifically for indian scripts, handwriting recognition has not been focused much till in the near past. However, due to generous Gove...
详细信息
Online handwriting recognition research has recently received significant thrust. Specifically for indian scripts, handwriting recognition has not been focused much till in the near past. However, due to generous Government funding through the group on Technology Development for indian Languages (TDIL) of the Ministry of Communication & Information Technology (MC&IT), Govt. of India, research in this area has received due attention and several groups are now engaged in research and development works for online handwriting recognition in different indian scripts. An extensive bottleneck of the desired progress in this area is the difficulty of collection of large sample databases of online handwriting in various scripts. Towards the same, recently a user-friendly tool on Android platform has been developed to collect data on handheld devices. This tool is called ISIgraphy and has been uploaded in the Google Play for free download. This application is designed well enough to store handwritten data samples in large scales in user-given file names for distinct users. Its use is script independent, meaning that it can collect and store handwriting samples written in any language, not necessarily an indian script. It has an additional module for retrieval and display of stored data. Moreover, it can directly send the collected data to others via electronic mail.
In this paper, we address the problem of separating the diffuse and specular reflection components of complex textured surfaces from a single color image. Unlike most previous approaches that assume accurate knowledge...
详细信息
In this paper, we address the problem of separating the diffuse and specular reflection components of complex textured surfaces from a single color image. Unlike most previous approaches that assume accurate knowledge of illumination source color for this task, we analyze errors in source color information to perform robust separation. The analysis leads to a simple, efficient and robust algorithm to estimate the diffuse and specular components using the estimated source color. The algorithm is completely automatic and does not need explicit color segmentation or color boundary detection as required by many existing methods. Results on complex textured images show the effectiveness of the proposed algorithm for robust reflection component separation.
We present an improved mesh denoising method based on 3D geometric bilateral filtering. Its novelty is that it can preserve the details of the object as well as reduce the noise in an effective manner. The previous ap...
详细信息
We present an improved mesh denoising method based on 3D geometric bilateral filtering. Its novelty is that it can preserve the details of the object as well as reduce the noise in an effective manner. The previous approach of geometric bilateral filtering for 3D-scan points has a limitation that it reduces the point density, thereby losing the details present in the object. The approach proposed by us, on the contrary, works on the surface mesh obtained after triangulating the 3D-scan points without any data downsampling. Each vertex of the mesh is repositioned appropriately based on the estimated centroid of the vertices in its local neighborhood and a Gaussian weight function. Experimental results demonstrate its strength, efficiency, and robustness.
The video coding standard H.264 uses Context-based Adaptive Variable Length Coding (CAVLC) as one of its entropy encoding techniques. This paper proposes VLSI architecture for CAVLC algorithm. The designed hardware me...
详细信息
The video coding standard H.264 uses Context-based Adaptive Variable Length Coding (CAVLC) as one of its entropy encoding techniques. This paper proposes VLSI architecture for CAVLC algorithm. The designed hardware meets the required speed of H.264 without compromising the hardware cost. The CAVLC encoder works at a maximum clock frequency of 126 MHz when implemented in Xilinx 10.1i, Virtex-5 technology. The speed is quite appreciable when compared to other existing works. The implemented architecture meets the required rate for processing of HD-1080 format video sequence.
In case of detection and analysis of deformation in collision scenarios, using a method based on area of contact instead of a point of contact generates numerically stable impulse forces. Area of contact improves the ...
详细信息
In case of detection and analysis of deformation in collision scenarios, using a method based on area of contact instead of a point of contact generates numerically stable impulse forces. Area of contact improves the stability of control algorithms, but it is often associated with high computational cost. In this paper, we alleviate this problem by proposing a novel algorithm for collision detection of a deformable mesh against rigid structures. We reuse the data structures maintained for elastic force computations in the FEM, for the purpose of collision detection. Parallel constructs on GPU using reduced model make the simulations interactive even for meshes with thousands of elements. Since we don't maintain any additional complex structure for keeping track of the deformable body at each iteration, we significantly reduce the usage of GPU memory bandwidth. Efficiency of our method is illustrated by reporting high culling efficiency on various tests.
Simultaneous measurement of multi-dimensional deformations can be done in digital holographic Moiré by illuminating the object with multiple beams from different angles and recording the interference pattern due ...
详细信息
Simultaneous measurement of multi-dimensional deformations can be done in digital holographic Moiré by illuminating the object with multiple beams from different angles and recording the interference pattern due to multiple beams. This technique involves reliable estimation of resultant multiple phase maps from a one record of the reconstructed interference moiré field. In this paper, we introduce a novel method, based on Unscented Kalman Filter (UKF), for the estimation of multiple phase maps from a complex multi-component signal encountered in digital holographic moiré. The State space model used in this method is determined by Taylor series expansion of phase component as process model and polar to Cartesian conversion as measurement model.
In this paper, we propose a modified version of the standard proportional-derivative (PD) controller for biped locomotion. Our improvements stabilize the biped for high gain PD controllers. The main idea of our approa...
详细信息
In this paper, we propose a modified version of the standard proportional-derivative (PD) controller for biped locomotion. Our improvements stabilize the biped for high gain PD controllers. The main idea of our approach involves applying corrective component to the existing framework, so that it prevents overshooting at high gains to stabilize the biped. We use pose control graphs to represent various gaits for the biped. We demonstrate with our improvements that the biped controller is stable while walking on irregular terrains. We also demonstrate that our formulation provides additional stability to the biped under minor impediments while in motion.
In this paper we present a max-flow min-cut based salient object detection in 3D point cloud that results from Structure from Motion (SfM) pipeline. The SfM pipeline generates noisy point cloud due to the unwanted sce...
详细信息
In this paper we present a max-flow min-cut based salient object detection in 3D point cloud that results from Structure from Motion (SfM) pipeline. The SfM pipeline generates noisy point cloud due to the unwanted scenes captured along with the object in the image dataset of SfM. The background points being sparse and not meaningful, it becomes necessary to remove them. Hence, any further processes (like surface reconstruction) utilizing the cleaned up model will have no hinderance from the noise removed. We present a novel approach where the camera centers are used to segment out the salient object. The algorithm is completely autonomous and does not need any user input. We test our proposed method on indian historical models reconstructed through SfM. We evaluate the results in terms of selectivity and specificity.
暂无评论