Skip Decision and Reference Frame Selection for Low-Complexity H.264/AVC Surveillance Video Coding

H. 264/advanced video coding surveillance video encoders use the Skip mode specified by the standard to reduce bandwidth. They also use multiple frames as reference for motion-compensated prediction. In this paper, we propose two techniques to reduce the bandwidth and computational cost of static camera surveillance video encoders without affecting detection and recognition performance. A spatial sampler is proposed to sample pixels that are segmented using a Gaussian mixture model. Modified weight updates are derived for the parameters of the mixture model to reduce floating point computations. A storage pattern of the parameters in memory is also modified to improve cache performance. Skip selection is performed using the segmentation results of the sampled pixels. The second contribution is a low computational cost algorithm to choose the reference frames. The proposed reference frame selection algorithm reduces the cost of coding uncovered background regions. We also study the number of reference frames required to achieve good coding efficiency. Distortion over foreground pixels is measured to quantify the performance of the proposed techniques. Experimental results show bit rate savings of up to 94.5% over methods proposed in literature on video surveillance data sets. The proposed techniques also provide up to 74.5% reduction in compression complexity without increasing the distortion over the foreground regions in the video sequence.

Coded Illumination for Motion-blur Free Imaging of Cells on Cell-phone Based Imaging Flow Cytometer

Cell-phone based imaging flow cytometry can be realized by flowing cells through the microfluidic devices, and capturing their images with an optically enhanced camera of the cell-phone. Throughput in flow cytometers is usually enhanced by increasing the flow rate of cells. However, maximum frame rate of camera system limits the achievable flow rate. Beyond this, the images become highly blurred due to motion-smear. We propose to address this issue with coded illumination, which enables recovery of high-fidelity images of cells far beyond their motion-blur limit. This paper presents simulation results of deblurring the synthetically generated cell/bead images under such coded illumination.

An Optical How Feature and McFIS Based Approach for 3-Dimensional Human Action Recognition

We propose to develop a 3-D optical flow features based human action recognition system. Optical flow based features are employed here since they can capture the apparent movement in object, by design. Moreover, they can represent information hierarchically from local pixel level to global object level. In this work, 3-D optical flow based features a re extracted by combining the 2-1) optical flow based features with the depth flow features obtained from depth camera. In order to develop an action recognition system, we employ a Meta-Cognitive Neuro-Fuzzy Inference System (McFIS). The m of McFIS is to find the decision boundary separating different classes based on their respective optical flow based features. McFIS consists of a neuro-fuzzy inference system (cognitive component) and a self-regulatory learning mechanism (meta-cognitive component). During the supervised learning, self-regulatory learning mechanism monitors the knowledge of the current sample with respect to the existing knowledge in the network and controls the learning by deciding on sample deletion, sample learning or sample reserve strategies. The performance of the proposed action recognition system was evaluated on a proprietary data set consisting of eight subjects. The performance evaluation with standard support vector machine classifier and extreme learning machine indicates improved performance of McFIS is recognizing actions based of 3-D optical flow based features.

Photosensitivity study of GeS2 chalcogenide glass under femtosecond laser pulses irradiation

The present study discusses the photosensitivity of GeS2 chalcogenide glass in response to irradiation with femtosecond pulses at 1047 nm. Bulk GeS2 glasses are prepared by conventional melt quenching technique and the amorphous nature of the glass is confirmed using X-ray diffraction. Ultrafast laser inscription technique is used to fabricate the straight channel waveguides in the glass. Single scan and multi scan waveguides are inscribed in GeS2 glasses of length 0.65 cm using a master oscillator power amplifier Yb doped fiber laser (IMRA mu jewel D400) with different pulse energy and translation speed. Diameters of the inscribed waveguides are measured and its dependence on the inscription parameters such as translation speed and pulse energy is studied. Butt coupling method is used to characterize the loss measurement of the inscribed optical waveguides. The mode field image of the waveguides is captured using CCD camera and compared with the mode field image of a standard SMF-28 fibers.

A novel interferometric technique to estimate thermal diffusivity of `optically transparent solid' using `Isothermal surface velocimetry'

One-dimensional transient heat flow is interpreted as a procession of `macro-scale translatory motion of indexed isothermal surfaces'. A new analytical model is proposed by introducing velocity of isothermal surface in Fourier heat diffusion equation. The velocity dependent function is extracted by revisiting `the concept of thermal layer of heat conduction in solid' and `exact solution' to estimate thermal diffusivity. The experimental approach involves establishment of 1 D unsteady heat flow inside the sample through Step-temperature excitation. A novel self-reference interferometer is utilized to separate a `unique isothermal surface' in time-varying temperature field. The translatory motion of the said isothermal surface is recorded using digital camera to estimate its velocity. From the knowledge of thermo-optic coefficient, temperature of the said isothermal surface is predicted. The performance of proposed method is evaluated for Quartz sample and compared with literature.

The Asian elephant is amongst the top three frugivores of two tree species with easily edible fruit

Large animal species are prone to local extirpation, but ecologists cannot yet predict how the loss of megaherbivores affects ecosystem processes such as seed dispersal. Few studies have compared the quantity and quality of seed dispersal by megaherbivores versus alternative frugivores in the wild, particularly for plant species with fruit easily consumed by many frugivorous species. In a disturbed tropical moist forest in India, we examine whether megaherbivores are a major frugivore of two tree species with easily edible, mammal-dispersed fruit. We quantify the relative fruit removal rates of Artocarpus chaplasha and Careya arborea, by the Asian elephant (Elephas maximus) and alternative dispersers. Through focal watches and camera trapping, we found the elephant to be amongst the top three frugivores for each tree species. Furthermore, seed transects under A. chaplasha show that arboreal frugivores discard seeds only a short distance from the parental tree, underscoring the elephant's role as a long-distance disperser. Our data provide unprecedented support for an old notion: megaherbivores may be key dispersers for a broad set of mammal-dispersed fruiting species, and not just fruit inaccessible to smaller frugivores. As such, the elephant may be particularly important for the functional ecology of the disturbed forests it still inhabits across tropical Asia.

PolyMage: Automatic Optimization for Image Processing Pipelines

This paper presents the design and implementation of PolyMage, a domain-specific language and compiler for image processing pipelines. An image processing pipeline can be viewed as a graph of interconnected stages which process images successively. Each stage typically performs one of point-wise, stencil, reduction or data-dependent operations on image pixels. Individual stages in a pipeline typically exhibit abundant data parallelism that can be exploited with relative ease. However, the stages also require high memory bandwidth preventing effective utilization of parallelism available on modern architectures. For applications that demand high performance, the traditional options are to use optimized libraries like OpenCV or to optimize manually. While using libraries precludes optimization across library routines, manual optimization accounting for both parallelism and locality is very tedious. The focus of our system, PolyMage, is on automatically generating high-performance implementations of image processing pipelines expressed in a high-level declarative language. Our optimization approach primarily relies on the transformation and code generation capabilities of the polyhedral compiler framework. To the best of our knowledge, this is the first model-driven compiler for image processing pipelines that performs complex fusion, tiling, and storage optimization automatically. Experimental results on a modern multicore system show that the performance achieved by our automatic approach is up to 1.81x better than that achieved through manual tuning in Halide, a state-of-the-art language and compiler for image processing pipelines. For a camera raw image processing pipeline, our performance is comparable to that of a hand-tuned implementation.

Experimental Investigations on Interaction of Two Drops by Thermocapillary-Buoyancy Migration

Experiments were performed, in a terrestrial environment, to study the migration and interaction of two drops with different diameters in matrix liquid under temperature gradient field. Pure soybean oil and silicon oil were used as matrix liquid and the drop liquid, respectively. The information on the motions of two drops was recorded by CCD camera system in the experiments to analyze the trajectories and velocities of the drops. Our experiments showed that, upon two drops approaching each other, the influence of the larger drop on the motion of the smaller one became significant. Meanwhile the smaller drop had a little influence on the larger one all the time. The oscillation of migration velocities of both drops was observed as they were approaching. For a short period the smaller drop even moved backward when it became side by side with the larger one during the migration. Although our experimental results on the behavior of two drops are basically consistent with the theoretical predictions, there are also apparent differences. 2006 Elsevier Ltd. All rights reserved. Keywords: Thermocapillary migration; Drop; Interaction; Oscillation 1. Introduction A bubble or drop will move when placed in another fluid with temperature gradient. This motion happens as a consequence of the variation of interfacial tension with temperature. Such a phenomenon is already known as Marangoni migration problem. With the development of microgravity science, bubble dynamics and droplet dynamics became a hot point problem of research because this investigation is very important for basic research as well as for applications in reduced gravity environment, such as space material science, chemical engineering and so on. Young et al. first investigated the thermocapillary migration of

Visualization of dust explosion under microgravity conditions

The objective of this work was to apply visualization methods to the experimental study of cornstarch dust-air mixture combustion in a closed vessel volume under microgravity conditions. A dispersion system with a small scale of turbulence was used in the experiments. A gas igniter initiated combustion of the dust-air mixture in the central or top part of the vessel. Flame propagation through the quiescent mixture was recorded by a high-speed video camera. Experiments showed a very irregular flame front and irregular distribution of the regions with local reactions of re-burning behind the flame front. at a later stage of combustion. Heat transfer from the hot combustion products to the walls is shown to have an important role in the combustion development. The maximum pressure and maximum rate of pressure rise were higher for flame propagation from the vessel center than for flame developed from the top pan of the vessel. The reason for smaller increase of the rate of pressure rise, for the flame developed from the top of the vessel. in comparison with that developed from the vessel center, was much faster increase of the contact surface of the combustion gases with the vessel walls. It was found that in dust flames only small part of hear was released at the flame front, the remaining part being released far behind it.

Underwater Acoustics and Cavitating Flow of Water Entry

The fluid mechanics of water entry is studied through investigating the underwater acoustics and the supercavitation. Underwater acoustic signals in water entry are extensively measured at about 30 different positions by using a PVDF needle hydrophone. From the measurements we obtain (1) the primary shock wave caused by the impact of the blunt body on free surface; (2) the vapor pressure inside the cavity; (3) the secondary shock wave caused by pulling away of the cavity from free surface; and so on. The supercavitation induced by the blunt body is observed by using a digital high-speed video camera as well as the single shot photography. The periodic and 3 dimensional motion of the supercavitation is revealed. The experiment is carried out at room temperature.

DPIV and Interferometry Technique for Velocity Field and Concentration Fields Measurement in Crystal Growth

The property of crystal depends seriously on the solution concentration distribution near the growth surface of a crystal. However, the concentration distributions are affected by the diffusion and convection of the solution. In the present experiment, the two methods of optical measurement are used to obtained velocity field and concentration field of NaClO₃ solution. The convection patterns in sodium chlorate (NaClO₃) crystal growth are measured by Digital Particle image Velocimetry (DPIV) technology. The 2-dimentional velocity distributions in the solution of NaClO3 are obtained from experiments. And concentration field are obtained by a Mach-Zehnder interferometer with a phase shift servo system. Interference patterns were recorded directly by a computer via a CCD camera. The evolution of velocity field and concentration field from dissolution to crystallization are visualized clearly. The structures of velocity fields were compared with that of concentration field.

Real-Time Measurement on Deformation Fields of Hole-Excavated Samples under Impact Tension

In this paper, the real-time deformation fields are observed in two different kinds of hole-excavated dog-bone samples loaded by an SHTB, including single hole sample and dual holes sample with the aperture size of 0.8mm. The testing system consists of a high-speed camera, a He-Ne laser, a frame grabber and a synchronization device with the controlling accuracy of I microsecond. Both the single hole expanding process and the interaction of the two holes are recorded with the time interval of 10 mu s. The observed images on the sample surface are analyzed by newly developed software based on digital correlation theory and a modified image processing method. The 2-D displacement fields in plane are obtained with a resolution of 50 mu m and an accuracy of 0.5 mu m. Experimental results obtained in this paper are proofed, by compared with FEM numerical simulations.

Real-Time Measurement on Deformation Fields of Notched Samples Under Impact Tension

In this paper, a real-time and in situ optical measuring system is reported to observe high-velocity deformations of samples subjected to impact loading. The system consists of a high-speed camera, a He-Ne laser, a frame grabber, a synchronization device and analysis software based on digital correlation theory. The optical system has been adapted to investigate the dynamic deformation field and its evolution in notched samples loaded by an split Hopkinson tension bar, with a resolution of 50 pin and an accuracy of 0.5 mum. Results obtained in experiments are discussed and compared with numerical simulations. It is shown that the measuring system is effective and valid.

Numerical Simulation of Pulsed Liquid Fluidized Bed and Its Experimental Validation

The pulsed liquid fluidized bed was studied using numerical simulation and experimental methods, The area-averaged two-fluid model (TFM) was used to simulate the pulsed fluidization. The bed expansion and collapse processes were simulated first and the phenomena obtained from the calculation were consistent with our previous experiments and observations. In the pulsed fluidization, the variation of bed height, the variations of particle velocity and concentration distribution were obtained and analyzed. Experiments were carried out to validate the simulation results. The pressure variation with time at different locations was measured using pressure transducers and compared with the simulated results. The variations of bed height and particle concentration distribution were recorded using a digital video camera recorder. The results were consistent with the simulation results as a whole.

Convective heat dissipation with lattice-frame materials

This paper presents experimental results on heat transfer and pressure drop for a compact heat sink made of fully triangulated, lightweight (porosity∼0.938), aluminum lattice-frame materials (LFMs). Due to the inherent structural anisotropy of the LFMs, two mutually perpendicular orientations were selected for the measurements. Constant heat flux was applied to the heat sink under steady state conditions, and dissipated by forced air convection. The experimental data were compared with those predicted from an analytical model based on fin analogy. The experimental results revealed that pressure drop is strongly dependent upon the orientation of the structure, due mainly to the flow blockage effect. For heat transfer measurements, typical local temperature distributions on the substrate under constant heat flux conditions were captured with infrared camera. The thermal behavior of LFMs was found to follow closely that of cylinder banks, with early transition Reynolds number (based on strut diameter) equal to about 300. The Nusselt number prediction from the fin-analogy correlates well with experimental measurements, except at low Reynolds numbers where a slightly underestimation is observed. Comparisons with empty channels and commonly used heat exchanger media show that the present LFM heat sink can remove heat approximately seven times more efficient than an empty channel and as efficient as a bank of cylinders at the same porosity level. The aluminum LFMs are extremely stiff and strong, making them ideal candidates for multifunctional structures requiring both heat dissipation and mechanical load carrying capabilities. © 2003 Elsevier Ltd. All rights reserved.