Compressive reconstruction algorithm based adaptive dictionary learning is developed and used for single-exposure non-scanning 3d imaging by interferenceless Coded Aperture Correlation Holography. Background noise and...
详细信息
An integral photography anddeconvolution techniques have been applied to observe plasmas, i.e. continuous translucent luminous objects. We experimentally succeeded in distinguishing the three-dimensional distribution...
详细信息
deep learning is nowadays a mature technique and it is widely applied to image processing and classification. In recent years, many authors tried to extend this approach also to 3d object classification. However, most...
详细信息
ISBN:
(纸本)9781728144610;9781728144603
deep learning is nowadays a mature technique and it is widely applied to image processing and classification. In recent years, many authors tried to extend this approach also to 3d object classification. However, most of the works in this field refers to complete models, while in many real applications the single acquisition with a vision system may only provide a partial object representation. Thus, the main goal of this work is to study the behaviour of classification neural networks when partial 3d models are considered. In particular, the analysis is focused on the classification reliability using partial point clouds, evaluating the influence of noise level and object scaling on the overall network performance. Tests are carried out both on synthetic point clouds, generated by simulation of common acquisition techniques, and on real clouds acquired by a Kinect device. This pushes towards the development of hybrid solutions, where training is made on simulated clouds and the testing takes place on real scanned objects, providing interesting suggestions for practical applications.
We overview the performance of three dimensional (3d) integral imaging based human gesture recognition techniques under degraded environments. Using 3d integral imaging-based strategies we find substantial improvement...
详细信息
The efficacy of dynamic image Analysis (dIA) for evaluating particle size and shape parameters was explored using three natural sands, having varying particle morphologies. Two-dimensional (2d) captures binary images ...
详细信息
The efficacy of dynamic image Analysis (dIA) for evaluating particle size and shape parameters was explored using three natural sands, having varying particle morphologies. Two-dimensional (2d) captures binary images of the particles as they free fall in the imaging frame. Although 2ddIA is practical for statistical size and shape analysis, there is a prevailing perception that it fails to fully quantify particle granulometry. In the past few years, 3ddIA has been introduced and has gained acceptance in the pharmaceutical industry. In 3ddIA the system tracks a particle as it falls through the imaging frame and captures gray-scale images from 8 to 12 perspectives of the same particle, and the results are analyzed using average values of these 2dimages, which are believed to verge on true particle morphology. Although 2d and3ddevices employ similar methodology they differ in resolution, frame rate, lighting systems, and algorithms. In this work we compare the performance of 2d and3ddIA. Particle size distributions were expressed using EQPC and a variety of Feret diameters, while particle shape descriptors including Aspect Ratio, Sphericity, Convexity and Roundness were compared for both systems. It is shown that 3ddIA requires a smaller number of sand particles to achieve mean particle shape values. Particle size characterization is generally independent of the machines and algorithms used in this study;however, 3ddIA provides maximum and minimum particle axes which are closer to the real sand particle sizes. image-based particle shape characterization is more sensitive to the technology employed;it largely depends on image quality, particle angularity, and a hierarchy of shape descriptors;thus, at this time shape analysis for engineering applications must be carried out with similar machines and algorithms. In particular, the image resolutions captured by the available 2d and3ddIA apparatus are 4 mu m and 15 mu m per pixel, respectively. At this time, the
We developed a Convolutional Neural Network to estimate depth on wide-angle images using panomorph lens with controlleddistortion. We simulated three different lens model and compared their performances based on thei...
详细信息
Optical see-through near-eye displays (OST-NEds) is a key component for the augmented reality (AR) applications. Although this technology has not yet permeated enough in our life as virtual reality headsets do, both a...
详细信息
We present fluorescence imaging of plant cells based on transport of intensity equation and Fresnel propagation. At first, the phase distribution is obtained from the recorded three-defocus fluorescence intensity imag...
详细信息
暂无评论