Parallel implementation of a real-time high dynamic range video system

Guthier, Benjamin; Kopf, Stephan; Wichtlhuber, Matthias; Effelsberg, Wolfgang

doi:10.3233/ica-130461

Cited by 9 publications

(5 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…And they are also available on common PC platforms. Therefore, from view of practice, we should try improve our algorithm with multicore/many-core acceleration platforms for industrial applications [14,27,77,78].…”

Section: Discussionmentioning

confidence: 99%

An efficient approach to directly compute the exact Hausdorff distance for 3D point sets

Zhang

Han

et al. 2017

ICA

View full text Add to dashboard Cite

Hausdorff distance measure is very important in CAD/CAE/CAM related applications. This manuscript presents an efficient framework and two complementary subalgorithms to directly compute the exact Hausdorff distance for general 3D point sets. The first algorithm of Nonoverlap Hausdorff Distance (NOHD) combines branch-and-bound with early breaking to cut down the Octree traversal time in case of spatial nonoverlap. The second algorithm of Overlap Hausdorff Distance (OHD) integrates a point culling strategy and nearest neighbor search to reduce the number of points traversed in case of spatial overlap. The two complementary subalgorithms can achieve a highly efficient and balanced result. Both NOHD and OHD compute the exact Hausdorff distance directly for arbitrary 3D point sets. We conduct a number of experiments on benchmark models and CAD application models, and compare the proposed approach with other state-of-the-art algorithms. The results demonstrate the effectiveness of our method.

show abstract

Section: Discussionmentioning

confidence: 99%

An efficient approach to directly compute the exact Hausdorff distance for 3D point sets

Zhang

Han

et al. 2017

ICA

View full text Add to dashboard Cite

show abstract

“…In the case of very complex 4D products, the computer processing time may be too high. As multi-core CPUs and GPUs become more available with higher performance and lower cost, this methodology could be accelerated [2,11,43]. Although SolidWorks was used due to the availability of the flex feature and the configurations module, more advanced FEA software such as Abaqus could also be investigated to allow the simulation of more complex parts and include acceleration techniques.…”

Section: Discussionmentioning

confidence: 99%

Lightweight parametric design optimization for 4D printed parts

Paz

Pei

Monzón

et al. 2017

ICA

View full text Add to dashboard Cite

4D printing is a technology that combines the capabilities of 3D printing with materials that can transform its geometry after being produced (e.g. Shape Memory Polymers). These advanced materials allow shape change by applying different stimulus such as heating. A 4D printed part will usually have 2 different shapes: a programmed shape (before the stimulus is applied), and the original shape (which is recovered once the stimulus has been applied). Lightweight parametric optimization techniques are used to find the best combination of design variables to reduce weight and lower manufacturing costs. However, current optimization techniques available in commercial 3D CAD software are not prepared for optimization of multiple shapes. The fundamental research question is how to optimize a design that will have different shapes with different boundary conditions and requirements. This paper presents a new lightweight parametric optimization method to solve this limitation. The method combines the Latin Hypercube design of experiments, Kriging metamodel and specifically designed genetic algorithms. The optimization strategy was implemented and automated using a CAD software. This method recognizes both shapes of the part as a single design and allows the lightweight parametric optimization to retain the minimum mechanical properties for both shapes.

show abstract

“…In particular, a recent popularized model known as Deep Learning and usually applied to large training data sets, relies in a training process that may take several days or even weeks to be completed [5,17]. In this sense alternatives based on cluster computing, GPUs and FPGAs are sensible strategies, each of them having their benefits and drawbacks [10,29,43,44]. In particular, Field Programmable Gate Arrays (FPGA) [18] are reprogrammable silicon chips, using prebuilt logic blocks and programmable routing resources that can be configured to implement custom hardware functionality.…”

Section: Introductionmentioning

confidence: 99%

Layer multiplexing FPGA implementation for deep back-propagation learning

Ortega-Zamorano

Jerez

Gómez

et al. 2017

ICA

View full text Add to dashboard Cite

Training of large scale neural networks, like those used nowadays in Deep Learning schemes, requires long computational times or the use of high performance computation solutions like those based on cluster computation, GPU boards, etc. As a possible alternative, in this work the Back-Propagation learning algorithm is implemented in an FPGA board using a multiplexing layer scheme, in which a single layer of neurons is physically implemented in parallel but can be reused any number of times in order to simulate multi-layer architectures. An on-chip implementation of the algorithm is carried out using a training/validation scheme in order to avoid overfitting effects. The hardware implementation is tested on several configurations, permitting to simulate architectures comprising up to 127 hidden layers with a maximum number of neurons in each layer of 60 neurons. We confirmed the correct implementation of the algorithm and compared the computational times against C and Matlab code executed in a multicore supercomputer, observing a clear advantage of the proposed FPGA scheme. The layer multiplexing scheme used provides a simple and flexible approach in comparison to standard implementations of the Back-Propagation algorithm representing an important step towards the FPGA implementation of deep neural networks, one of the most novel and successful existing models for prediction problems.

show abstract

Parallel implementation of a real-time high dynamic range video system

Cited by 9 publications

References 24 publications

An efficient approach to directly compute the exact Hausdorff distance for 3D point sets

An efficient approach to directly compute the exact Hausdorff distance for 3D point sets

Lightweight parametric design optimization for 4D printed parts

Layer multiplexing FPGA implementation for deep back-propagation learning

Contact Info

Product

Resources

About