Visual quality optimization for real-time encoders

Ambient occlusion (AO) is a popular rendering technique that enhances depth perception and realism by darkening locations that are less exposed to ambient light (e.g., corners and creases). In real-time applications, screen-space variants, relying on the depth buffer, are used due to their high performance and good visual quality. However, these only take visible surfaces into account, resulting in inconsistencies, especially during motion. Stochastic-Depth Ambient Occlusion is a novel AO algorithm that accounts for occluded geometry by relying on a stochastic depth map, capturing multiple scene layers per pixel at random. Hereby, we efficiently gather missing information in order to improve upon the accuracy and spatial stability of conventional screen-space approximations, while maintaining real-time performance. Our approach integrates well into existing rendering pipelines and improves the robustness of many different AO techniques, including multi-view solutions.

Download Full-text

Contrast Enhancement and Optimization for Image with Fog

Applied Mechanics and Materials ◽

10.4028/www.scientific.net/amm.380-384.3778 ◽

2013 ◽

Vol 380-384 ◽

pp. 3778-3781

Author(s):

Wei Na Huang ◽

Zheng Xiang Xie

Keyword(s):

Contrast Enhancement ◽

Real Time ◽

Visual Quality ◽

Experimental Results ◽

Video Monitoring ◽

Adaptive Model ◽

Absorption Effect ◽

Atmospheric Visibility ◽

Normal Work ◽

Computing Speed

Aiming at the absorption effect of fog suspended in the atmosphere on light, the paper established the removing-fog compensation adaptive model which can improve the atmospheric visibility and restore the normal work of outdoor system. The experimental results show that the removing fog image processed by the method of removing-fog compensation optimization can accord with the requirement of human visual, and it can be used in real-time video monitoring as the fast computing speed. The method not only can be used in foggy video which the fog distributed uniformly, and can assess the visual quality for the images processed.

Download Full-text

On Optimizing the Visual Quality of HASM-Based Streaming—The Study the Sensitivity of Motion Estimation Techniques for Mesh-Based Codecs in Ultra High Definition Large Format Real-Time Video Coding

Enabling Machine Learning Applications in Data Science - Algorithms for Intelligent Systems ◽

10.1007/978-981-33-6129-4_15 ◽

2021 ◽

pp. 207-219

Author(s):

Khaled Ezzat ◽

Ahmed Tarek Mohamed ◽

Ibrahim El-Shal ◽

Wael Badawy

Keyword(s):

Motion Estimation ◽

Video Coding ◽

Real Time ◽

Visual Quality ◽

High Definition ◽

Large Format ◽

Estimation Techniques

Download Full-text

A Real-Time Visualization Method of Objects Using Grid Coding Based on Surface Orientation and Structure

Advanced Materials Research ◽

10.4028/www.scientific.net/amr.462.3 ◽

2012 ◽

Vol 462 ◽

pp. 3-9

Author(s):

Lin Xu ◽

Sheng Yuan Yan

Keyword(s):

Real Time ◽

Systematic Approach ◽

Visual Quality ◽

Parametric Surface ◽

Excellent Performance ◽

Central Difference ◽

Visible Object ◽

Real Time Visualization ◽

Real Time Applications ◽

Virtual Reality Application

A systematic approach is introduced to infer surface structure of visible object surfaces using grid coding in this paper. The visualized object, which has surface of a terrain-like shape, is assumed to derive from data field of gird points defined by orthogonal division and data acquisition. Parametric surface is used to spatially represent the data for analysis. Central difference technology is presented to tessellate the curved surface and generate view-dependent multi-resolution models, which has excellent performance in real-time applications. Besides, an algorithm for data pre-processing is presented. The approach presented here also adapts to data acquired in realistic scene, which can be a virtual reality application. Comparing to previous methods, the method presented in this paper provides good flexibility in various applications. It works well in real-time graphic application while has good visual quality.

Download Full-text

Pengaruh Konversi Nurbs Ke Polygonal Pada Desain Mobil 3d Terhadap Penilaian Kualitas 3d Model

Creative Information Technology Journal ◽

10.24076/citec.2019v6i2.250 ◽

2021 ◽

Vol 6 (2) ◽

pp. 119

Author(s):

Awaludin Abid ◽

Kusrini Kusrini ◽

Amir Fatah Sofyan

Keyword(s):

Virtual Reality ◽

Augmented Reality ◽

Real Time ◽

Hausdorff Distance ◽

Visual Quality ◽

Quality Metrics ◽

Data Preparation ◽

Mesh Quality ◽

Output Data ◽

3D Mesh

Di Industri otomotif, biaya prototyping meningkat berbanding lurus dengan kompleksitas dan dependensi kendaraan. Sebagai alternatif untuk prototyping fisik dapat memanfaatkan teknologi baru seperti Augmented Reality (AR) dan Virtual Reality (VR) digunakan. Penggunaan VR dan AR melibatkan real-time rendering data CAD yang mengkonsumsi banyak memori dan mengurangi kinerja aplikasi. Persiapan data memiliki peran penting untuk meningkatkan kinerja sementara tetap mempertahankan topologi dan kualitas mesh. Proses optimalisasi data CAD yang digunakan yaitu Tessellation atau mengkonversi NURBS ke Polygons, berperan untuk menghasilkan output data yang memiliki efisien kinerja dengan topologi serta kualitas mesh yang baik. Hadirnya software 3D Data preparation dan optimasi pada kelas Tessellator. Autodesk Maya merupakan software pemodelan 3D yang mendukung Non-Uniform Rational Basis Spline ataupun CAD memiliki fitur mengkonversi model NURBS ke polygons, pemilihan kebutuhan atau requirement pada tessellation berpengaruh terhadap hasil output. Penilaian dilakukan menggunakan penilaian Objektif menggunakan 3D mesh visual quality metrics berbasis vertex-position Hausdorff Distance sehingga didapatkan requirement pada Tessellation yang efektif. Hasil dari konversi memiliki topologi yang serupa dengan software khusus data preparation dan optimasi, sedangkan hasil penilaian mesh visual quality metrics requirement yang mendekati yaitu menggunakan Tessellation Method Count dan General. Kata Kunci— Tessellation, Mesh Visual Quality, CAD, Polygon In automotive industry, cost of prototyping increases directly with complexity and dependencies of vehicle. As an alternative to physical prototyping can utilize new technologies such as Augmented Reality (AR) and Virtual Reality (VR) are used. And involves the real-time rendering of CAD data which consumes a lot of memory and reduces application performance. Data preparation has an important role to improve performance while maintaining topology and mesh quality. Process of optimizing CAD data used is Tessellation or converting NURBS to Polygons, whose role is to produce output data that has an efficient performance with topology and good mesh quality. Autodesk Maya is a 3D modeling software that supports Non-Uniform Rational Base Spline or CAD which has the feature of converting NURBS models to polygons, the selection of requirements or requirements on tessellation influences the output results. The assessment is done using objective assessment with 3D mesh visual quality metrics based on Hausdorff Distance vertex-position so that the requirements for effective Tessellation are obtained. The results of the conversion have a topology similar to special data preparation and optimization software, while the results of the mesh visual quality metrics requirement approach are close to using the Count and General Tessellation method. Keywords— Tessellation, Mesh Visual Quality, CAD, Polygon

Download Full-text

Responses of Landscape Groundcovers to Minimum Irrigation

Journal of Environmental Horticulture ◽

10.24266/0738-2898-19.2.78 ◽

2001 ◽

Vol 19 (2) ◽

pp. 78-84 ◽

Cited By ~ 7

Author(s):

Dennis R. Pittenger ◽

David A. Shaw ◽

Donald R. Hodel ◽

Douglas B. Holt

Keyword(s):

Real Time ◽

Reference Evapotranspiration ◽

Irrigation Treatment ◽

Visual Quality ◽

Hedera Helix ◽

Applied Water ◽

Twin Peaks ◽

Baccharis Pilularis ◽

Vinca Major

Abstract Four irrigation regimes based on percentages of real-time reference evapotranspiration (ET0) were applied to six widely used species of landscape groundcovers (Baccharis pilularis ‘Twin Peaks’, Drosanthemum hispidum, Vinca major, Gazania rigens v. leucolaena ‘Yellow Cascade’, Potentilla tabernaemontanii, and Hedera helix ‘Needlepoint’) during a 17-month period in Irvine, CA. Irrigation treatments (50%, 40%, 30% and 20% of ET0) were applied when accumulated real-time ET0 × treatment percentage reached 4.0 cm (1.6 in). Although the response to irrigation treatment was species dependent, Baccharis, Drosanthemum, and Hedera maintained at least minimally acceptable visual quality with applied water equal to 20% ET0, while Vinca required a minimum of 30% ET0. Acceptable visual quality of Gazania and Potentilla were not maintained at any treatment. Visual quality of Potentilla was better at ET0 greater than or equal to 30% ET0, but visual quality of Gazania was not improved with more water.

Download Full-text

Real-Time Large Crowd Rendering with Efficient Character and Instance Management on GPU

International Journal of Computer Games Technology ◽

10.1155/2019/1792304 ◽

2019 ◽

Vol 2019 ◽

pp. 1-15

Author(s):

Yangzi Dong ◽

Chao Peng

Keyword(s):

Real Time ◽

Visual Quality ◽

Level Of Detail ◽

Experimental Results ◽

Visual Appearance ◽

Fine Grained ◽

Individual Character ◽

Visibility Culling ◽

Rendering Techniques

Achieving the efficient rendering of a large animated crowd with realistic visual appearance is a challenging task when players interact with a complex game scene. We present a real-time crowd rendering system that efficiently manages multiple types of character data on the GPU and integrates seamlessly with level-of-detail and visibility culling techniques. The character data, including vertices, triangles, vertex normals, texture coordinates, skeletons, and skinning weights, are stored as either buffer objects or textures in accordance with their access requirements at the rendering stage. Our system preserves the view-dependent visual appearance of individual character instances in the crowd and is executed with a fine-grained parallelization scheme. We compare our approach with the existing crowd rendering techniques. The experimental results show that our approach achieves better rendering performance and visual quality. Our approach is able to render a large crowd composed of tens of thousands of animated instances in real time by managing each type of character data in a single buffer object.

Download Full-text