We present an approximate volume rendering algorithm that can compute multiple views of a 3D voxel-based data set concurrently. The approach employs a unique new method for combining partial results from neighboring objections to compute a sequence of rotated views, in fewer instructions than would be required for independent computations. For instance, the algorithm can compute a set of N projections through an N/spl times/N/spl times/N data set in only O(log N) parallel steps, using only O(N/sup 3/) total operations (work), matching the bounds for computing a single projection by conventional methods.
{"title":"Parallel approximate computation of projections for animated volume rendered displays","authors":"Tung-Kuang Wu, M. Brady","doi":"10.1145/166181.166190","DOIUrl":"https://doi.org/10.1145/166181.166190","url":null,"abstract":"We present an approximate volume rendering algorithm that can compute multiple views of a 3D voxel-based data set concurrently. The approach employs a unique new method for combining partial results from neighboring objections to compute a sequence of rotated views, in fewer instructions than would be required for independent computations. For instance, the algorithm can compute a set of N projections through an N/spl times/N/spl times/N data set in only O(log N) parallel steps, using only O(N/sup 3/) total operations (work), matching the bounds for computing a single projection by conventional methods.","PeriodicalId":394370,"journal":{"name":"Proceedings of 1993 IEEE Parallel Rendering Symposium","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1993-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129516249","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Interactive volume rendering is important to the timely analysis of three-dimensional data, but workstations take seconds to minutes to render data sets of a few megabytes. We have developed a parallel ray-casting technique, called Segmented Ray Casting, which can render a 128/spl times/128/spl times/128 data set at 2-3 frames per second on a 4K processor DECmpp 1200/Sx Model 100. Pixel values in the image plane are computed by casting rays through the volume data. The rays are segmented based on the intersection with the data sublocks in the processors. Each processor computes the color and opacity of the ray segments which pass through its subblock, which are then sent to the appropriate processor for composition with other segment values. Unlike other data-parallel volume renderers, Segmented Ray Casting does not require the transposition of volume data between processors at any time, nor does it suffer from resampling artifacts due to shearing.
交互式体绘制对于三维数据的及时分析非常重要，但是工作站需要几秒到几分钟才能呈现几兆字节的数据集。我们已经开发了一种并行光线投射技术，称为分段光线投射，它可以在4K处理器DECmpp 1200/Sx Model 100上以每秒2-3帧的速度渲染128/spl次/128/spl次/128数据集。图像平面中的像素值是通过通过体数据投射光线来计算的。射线是根据与处理器中的数据子锁的交集来分割的。每个处理器计算通过其子块的光线段的颜色和不透明度，然后将其发送到适当的处理器与其他段值组合。与其他数据并行体渲染器不同，分段光线投射不需要在任何时候在处理器之间转换体数据，也不会由于剪切而遭受重采样伪影。
{"title":"Segmented ray casting for data parallel volume rendering","authors":"William M. Hsu","doi":"10.1145/166181.166182","DOIUrl":"https://doi.org/10.1145/166181.166182","url":null,"abstract":"Interactive volume rendering is important to the timely analysis of three-dimensional data, but workstations take seconds to minutes to render data sets of a few megabytes. We have developed a parallel ray-casting technique, called Segmented Ray Casting, which can render a 128/spl times/128/spl times/128 data set at 2-3 frames per second on a 4K processor DECmpp 1200/Sx Model 100. Pixel values in the image plane are computed by casting rays through the volume data. The rays are segmented based on the intersection with the data sublocks in the processors. Each processor computes the color and opacity of the ray segments which pass through its subblock, which are then sent to the appropriate processor for composition with other segment values. Unlike other data-parallel volume renderers, Segmented Ray Casting does not require the transposition of volume data between processors at any time, nor does it suffer from resampling artifacts due to shearing.","PeriodicalId":394370,"journal":{"name":"Proceedings of 1993 IEEE Parallel Rendering Symposium","volume":"37 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1993-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127614155","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The two key issues in implementing a parallel ray-casting volume renderer are the work distribution and the data distribution. We have implemented such a renderer on the Fujitsu AP1000 using an adaptive image-space subdivision algorithm based on the worker-farm paradigm for the work distribution, and a distributed virtual memory, implemented in software, to provide the data distribution. Measurements show that this scheme works efficiently and effectively utilizes the data coherence that is inherent in volume data.
{"title":"Parallel volume rendering and data coherence","authors":"B. Corrie, P. Mackerras","doi":"10.1145/166181.166184","DOIUrl":"https://doi.org/10.1145/166181.166184","url":null,"abstract":"The two key issues in implementing a parallel ray-casting volume renderer are the work distribution and the data distribution. We have implemented such a renderer on the Fujitsu AP1000 using an adaptive image-space subdivision algorithm based on the worker-farm paradigm for the work distribution, and a distributed virtual memory, implemented in software, to provide the data distribution. Measurements show that this scheme works efficiently and effectively utilizes the data coherence that is inherent in volume data.","PeriodicalId":394370,"journal":{"name":"Proceedings of 1993 IEEE Parallel Rendering Symposium","volume":"5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1993-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123905009","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This paper describes a multiresolution approach to the visualization of surface data. The algorithms discussed allow the generation of arbitrary views of 3-dimensional surfaces. Image processing and texture mapping techniques are combined in a new 3-pass scanline algorithm to achieve smooth and continuous translations, rotations, and scale changes of large data sets. The implementation of the algorithms on a massively parallel SIMD video supercomputer, the Princeton Engine, allows the scenes to be generated interactively at video rates.
{"title":"A pyramid-based approach to interactive terrain visualization","authors":"James K. Tam, J. Peters","doi":"10.1145/166181.166191","DOIUrl":"https://doi.org/10.1145/166181.166191","url":null,"abstract":"This paper describes a multiresolution approach to the visualization of surface data. The algorithms discussed allow the generation of arbitrary views of 3-dimensional surfaces. Image processing and texture mapping techniques are combined in a new 3-pass scanline algorithm to achieve smooth and continuous translations, rotations, and scale changes of large data sets. The implementation of the algorithms on a massively parallel SIMD video supercomputer, the Princeton Engine, allows the scenes to be generated interactively at video rates.","PeriodicalId":394370,"journal":{"name":"Proceedings of 1993 IEEE Parallel Rendering Symposium","volume":"64 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1993-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126747795","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Volume rendering algorithms visualize sampled three dimensional data. A variety of applications create sampled data, including medical imaging, simulations, animation, and remote sensing. Researchers have sought to speed up volume rendering because of the high run time and wide application. Our algorithm uses permutation warping to achieve linear speedup on data parallel machines. This new algorithm calculates higher quality images than previous distributed approaches, and also provides more view angle freedom. We present permutation warping results on the SIMD MasPar MP-1. The efficiency results from nonconflicting communication. The communication remains efficient with arbitrary view directions, larger data sets, larger parallel machines, and high order filters. We show constant run time versus view angle, tunable filter quality, and efficient memory implementation.
{"title":"Permutation warping for data parallel volume rendering","authors":"C. Wittenbrink, Arun Kumar Somani","doi":"10.1145/166181.166189","DOIUrl":"https://doi.org/10.1145/166181.166189","url":null,"abstract":"Volume rendering algorithms visualize sampled three dimensional data. A variety of applications create sampled data, including medical imaging, simulations, animation, and remote sensing. Researchers have sought to speed up volume rendering because of the high run time and wide application. Our algorithm uses permutation warping to achieve linear speedup on data parallel machines. This new algorithm calculates higher quality images than previous distributed approaches, and also provides more view angle freedom. We present permutation warping results on the SIMD MasPar MP-1. The efficiency results from nonconflicting communication. The communication remains efficient with arbitrary view directions, larger data sets, larger parallel machines, and high order filters. We show constant run time versus view angle, tunable filter quality, and efficient memory implementation.","PeriodicalId":394370,"journal":{"name":"Proceedings of 1993 IEEE Parallel Rendering Symposium","volume":"98 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1993-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133962567","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：