Showing 1–6 of 6 results for author: Chua, N

SpiralMLP: A Lightweight Vision MLP Architecture

Authors: Haojie Mu, Burhan Ul Tayyab, Nicholas Chua

Abstract: We present SpiralMLP, a novel architecture that introduces a Spiral FC layer as a replacement for the conventional Token Mixing approach. Differing from several existing MLP-based models that primarily emphasize axes, our Spiral FC layer is designed as a deformable convolution layer with spiral-like offsets. We further adapt Spiral FC into two variants: Self-Spiral FC and Cross-Spiral FC, which en… ▽ More We present SpiralMLP, a novel architecture that introduces a Spiral FC layer as a replacement for the conventional Token Mixing approach. Differing from several existing MLP-based models that primarily emphasize axes, our Spiral FC layer is designed as a deformable convolution layer with spiral-like offsets. We further adapt Spiral FC into two variants: Self-Spiral FC and Cross-Spiral FC, which enable both local and global feature integration seamlessly, eliminating the need for additional processing steps. To thoroughly investigate the effectiveness of the spiral-like offsets and validate our design, we conduct ablation studies and explore optimal configurations. In empirical tests, SpiralMLP reaches state-of-the-art performance, similar to Transformers, CNNs, and other MLPs, benchmarking on ImageNet-1k, COCO and ADE20K. SpiralMLP still maintains linear computational complexity O(HW) and is compatible with varying input image resolutions. Our study reveals that targeting the full receptive field is not essential for achieving high performance, instead, adopting a refined approach offers better results. △ Less

Submitted 31 March, 2024; originally announced April 2024.

RTSDF: Real-time Signed Distance Fields for Soft Shadow Approximation in Games

Authors: Yu Wei Tan, Nicholas Chua, Clarence Koh, Anand Bhojan

Abstract: Signed distance fields (SDFs) are a form of surface representation widely used in computer graphics, having applications in rendering, collision detection and modelling. In interactive media such as games, high-resolution SDFs are commonly produced offline and subsequently loaded into the application, representing rigid meshes only. This work develops a novel technique that combines jump flooding… ▽ More Signed distance fields (SDFs) are a form of surface representation widely used in computer graphics, having applications in rendering, collision detection and modelling. In interactive media such as games, high-resolution SDFs are commonly produced offline and subsequently loaded into the application, representing rigid meshes only. This work develops a novel technique that combines jump flooding and ray tracing to generate approximate SDFs in real-time. Our approach can produce relatively accurate scene representation for rendering soft shadows while maintaining interactive frame rates. We extend our previous work with details on the design and implementation as well as visual quality and performance evaluation of the technique. △ Less

Submitted 11 October, 2022; originally announced October 2022.

ACM Class: I.3

A Hybrid System for Real-time Rendering of Depth of Field Effect in Games

Authors: Yu Wei Tan, Nicholas Chua, Nathan Biette, Anand Bhojan

Abstract: Real-time depth of field in game cinematics tends to approximate the semi-transparent silhouettes of out-of-focus objects through post-processing techniques. We leverage ray tracing hardware acceleration and spatio-temporal reconstruction to improve the realism of such semi-transparent regions through hybrid rendering, while maintaining interactive frame rates for immersive gaming. This paper exte… ▽ More Real-time depth of field in game cinematics tends to approximate the semi-transparent silhouettes of out-of-focus objects through post-processing techniques. We leverage ray tracing hardware acceleration and spatio-temporal reconstruction to improve the realism of such semi-transparent regions through hybrid rendering, while maintaining interactive frame rates for immersive gaming. This paper extends our previous work with a complete presentation of our technique and details on its design, implementation, and future work. △ Less

Submitted 11 October, 2022; originally announced October 2022.

Comments: Best Student Paper Award (GRAPP 2022)

ACM Class: I.3

Hybrid DoF: Ray-Traced and Post-Processed Hybrid Depth of Field Effect for Real-Time Rendering

Authors: Yu Wei Tan, Nicholas Chua, Nathan Biette, Anand Bhojan

Abstract: Depth of Field (DoF) in games is usually achieved as a post-process effect by blurring pixels in the sharp rasterized image based on the defined focus plane. This paper describes a novel real-time DoF technique that uses ray tracing with image filtering to achieve more accurate partial occlusion semi-transparencies on edges of blurry foreground geometry. This hybrid rendering technique leverages r… ▽ More Depth of Field (DoF) in games is usually achieved as a post-process effect by blurring pixels in the sharp rasterized image based on the defined focus plane. This paper describes a novel real-time DoF technique that uses ray tracing with image filtering to achieve more accurate partial occlusion semi-transparencies on edges of blurry foreground geometry. This hybrid rendering technique leverages ray tracing hardware acceleration as well as spatio-temporal reconstruction techniques to achieve interactive frame rates. △ Less

Submitted 10 October, 2022; originally announced October 2022.

ACM Class: I.3

Journal ref: ACM SIGGRAPH 2020 Posters

RTSDF: Generating Signed Distance Fields in Real Time for Soft Shadow Rendering

Authors: Yu Wei Tan, Nicholas Chua, Clarence Koh, Anand Bhojan

Abstract: Signed Distance Fields (SDFs) for surface representation are commonly generated offline and subsequently loaded into interactive applications like games. Since they are not updated every frame, they only provide a rigid surface representation. While there are methods to generate them quickly on GPU, the efficiency of these approaches is limited at high resolutions. This paper showcases a novel tec… ▽ More Signed Distance Fields (SDFs) for surface representation are commonly generated offline and subsequently loaded into interactive applications like games. Since they are not updated every frame, they only provide a rigid surface representation. While there are methods to generate them quickly on GPU, the efficiency of these approaches is limited at high resolutions. This paper showcases a novel technique that combines jump flooding and ray tracing to generate approximate SDFs in real-time for soft shadow approximation, achieving prominent shadow penumbras while maintaining interactive frame rates. △ Less

Submitted 10 October, 2022; originally announced October 2022.

ACM Class: I.3

Journal ref: Pacific Graphics Short Papers, Posters, and Work-in-Progress Papers (2020)

Pre-Training Transformers for Domain Adaptation

Authors: Burhan Ul Tayyab, Nicholas Chua

Abstract: The Visual Domain Adaptation Challenge 2021 called for unsupervised domain adaptation methods that could improve the performance of models by transferring the knowledge obtained from source datasets to out-of-distribution target datasets. In this paper, we utilize BeiT [1] and demonstrate its capability of capturing key attributes from source datasets and apply it to target datasets in a semi-supe… ▽ More The Visual Domain Adaptation Challenge 2021 called for unsupervised domain adaptation methods that could improve the performance of models by transferring the knowledge obtained from source datasets to out-of-distribution target datasets. In this paper, we utilize BeiT [1] and demonstrate its capability of capturing key attributes from source datasets and apply it to target datasets in a semi-supervised manner. Our method was able to outperform current state-of-the-art (SoTA) techniques and was able to achieve 1st place on the ViSDA Domain Adaptation Challenge with ACC of 56.29% and AUROC of 69.79%. △ Less

Submitted 18 December, 2021; originally announced December 2021.