1,721,161 research outputs found
Parallel digital processing system for video and computer graphics applications
No full textThis project was carried out between February 1998 and January 2002 and involved a budget of $557,801 from the Academic Research Fund under reference number ARC 8/97. Research was carried out under the supervision of the principal investigator, Dr Seah Hock Soon and collaborators, Dr Lau Chiew Tong, Dr Cai Wentong and Dr Chan Syin. The purpose is to investigate novel digital image processing methods with the objective of utilizing them in video and computer graphics applications. As digital image processes are compute intensive, issues of parallel processing are also investigated
Deep reinforcement learning for table tennis
No full textDeep Reinforcement Learning (DRL) is a powerful branch of machine learning that
combines deep learning with reinforcement learning (RL) to enable intelligent agents
to learn complex behaviors. DRL has achieved remarkable success in areas such as
robotics, game playing (e.g., AlphaGo, OpenAI Five), autonomous vehicles, and
finance.
This report compares various configurations of DRL algorithms to train agents in
games. This report also continues with the work by Professor Seah Hock Soon and
his team, who built a novel table tennis game in the Unity environment and
successfully trained agents of different proportions to perform forehand shots using
DRL.
We attempted to teach an agent more complicated tasks. Firstly, we experimented
with backhand shots before attempting to teach an agent to perform both forehand
and backhand shots appropriately. Through these attempts, we discovered more
methods to improve results and training with DRL.
The key contribution of this report is the discovery of methods to improve training
speed and effectiveness DRL training, especially when rewards are sparse, like in
the table tennis environment, which can be valuable when DRL is applied to other
areas.Bachelor's degre
3D ultrasound imaging in medicine
No full textThis research deals with the problem of three dimensional reconstruction from medical ultrasound images. Based on an arbitrary data set of two dimensional ultrasound images, a system has been successfully developed to reconstruct and display 3D ultrasound images. The reconstruction task involves images acquisition from ultrasound machine, image processing, data set manipulation and rendering the volume data set to display image in three-dimension.Master of Applied Scienc
A shading effects framework for comics
No full textShading is a process in which shades of objects are drawn with different levels of darkness to create lifelike drawings. As technology advances, comic drawing is created gradually on the internet platform. Therefore, a need for creating quality shading effect for computer artworks was identified for this project.
In this project, a shading effects framework was developed using C++ programming language and Qt cross-platform application framework. The objective of this framework is to provide the ability for comic artists to create shading-like colour gradient for two-dimensional polygons. The colour gradient is created through the use of colour interpolation where colour interpolation can be defined as finding and creating new colour and coordinates of new points based on a given set of known points.
Information and references related to colour interpolation and colour space were sourced from articles, papers and books. Algorithms such as linear interpolation, conical interpolation and scan-line fill were identified and discussed. Linear interpolation and scan-line fill algorithm were chosen to implement the interpolation of colour in this project. Two colour spaces, the red-green-blue colour space and hue-saturation-value colour space were selected.Bachelor of Engineering (Computer Science
Softcell constraints for robust real-time deformation
No full textThis thesis presents softcell constraints, a novel set of formulations for robust real-time deformation. The underlying model handles volumetric objects in the form of surface mesh. The main theory behind the approach is to voxelize the surface mesh into cuboid cells. Neighbouring cells are elastically coupled together through constraint forces and torques. Based on geometric formulations, the constraint forces/torques are designed to minimize the global elastic energy. The rigidity of the cells is inspired by the coupled prisms method and has made the method highly robust even when under extreme deformation. The simulation cost is linear to the number of cells. Thus, it is capable of handling a reasonable number of complex objects. Collisions are easy to handle as it is based on cuboid primitives. Thus, it eliminates the burden to implement complex collision detection technique and contact handling. This combination of robustness, performance and ease of implementation makes the method particularly interesting for games.MASTER OF ENGINEERING (SCE
Ray tracing acceleration structure : GPU based KD-tree using SAH heuristic
No full textRay tracing is a very widely used rendering technique in many applications in the movie industry,
architecture and media industry. The reason why ray tracing is so popular is because it can produce the
most realistic and most accurate images compared to other rendering techniques due to its straight
forward nature of emulating light. However computational costs prevent it from being widely used in
interactive applications.
Much research has been done to help accelerate the ray tracing process and one of the most popular data
structures being used is the KD-Tree. However KD-Tree construction is also a computationally expensive
task thus holding it back from being used in interactive applications in which the scenes contain
animations and require the tree to be rebuilt in every frame.
With the advance of multi-purpose graphics hardware with its massive parallelism it is now possible to
accelerate the ray tracing process further to improve the performance greatly from traditional CPU based
approaches. The GPU construction stage also benefits greatly from the computing power of modern
graphics hardware and thus it is now possible to construct KD-Tree in real time.
In this final year project we try to implement such an approach by making a n early full GPU based ray
tracer using KD-Tree as the acceleration structure. The whole KD-Tree construction algorithm is also
implemented to try to take advantage of the GPU as much as possible using the CUDA framework
running on NVIDIA graphics cards.
Due to time constraints our GPU ray tracer is quite simple and does not support animation, however the
code is functional and reasonably stable. Further optimizations can be made to improve the performance
and more features can be implemented in future projects.Bachelor of Engineering (Computer Science
GPU acceleration
No full textOther than graphics processing abilities, GPU is widely used for general purpose computing nowadays because of its high parallel processing abilities. However, in order to utilize GPU as computation resources, there are prerequisites that the software needs to be programmed in a parallel manner which is more complex and expensive than in a serial manner. Many existing commercial video converter tools utilize CPU as the only computational resource for the video transcoding process. But occasionally the process can be very time-consuming, depending on the video size and the image quality that the user is pursuing. By bringing in GPU as a hardware-accelerated decoder, the whole process can be speedup by many times with lesser power consumed. As thread computing emerge, many in the industry start to shift the intensive computation part in video transcoding process from CPU and GPU, which brings a lot of performance improvement to the process. Everything seems perfect with the GPU acceleration transcoding until the rise of the dispute about the resulting image quality from a CPU-only software decoder and GPU-accelerated hardware decoder. Therefore this project will go deep into the basic concepts of GPU and video transcoding and examine the suitability for GPU as a hardware-accelerated decoder for video transcoding process.Bachelor of Engineering (Computer Engineering
Design and development of a native client application for mobile platform
No full textWith telecommuting becoming a more popular option, many organizations are using asset management software to maximize the advantages of telecommuting. Mobile devices have also become an integral part of daily lives, hence a mobile application for asset management can make mobile devices an important tool for telecommuting.
The aim of this project to design and develop a mobile application for an online asset management system called Asset Management and Production System (AMPS) that is designed for different media creation industries on the iOS platform.
The main focus for the mobile application is to develop a login interface for the system, a viewing interface for users to view the projects and the folders and assets inside the project and also allow some interactions with the application to accomplish tasks like managing the assets.
During the duration of the project, the basic functionalities were implemented in the application. These functionalities include the authentication of the user, displaying of the folders and assets in the project correctly, managing the assets, such as downloading, uploading, deletion and preview of the assets.
The report also discusses the possibility of improving the application in the future by adding more functionalities and improving the user interface.Bachelor of Engineering (Computer Engineering
3D face mapping via Kinect
No full textWhen modeling 3D faces from 2D images, many different views of the same stationary face model are required. 2D images are also more susceptible to interference from the environment. The Kinect sensor has been shown to provide reliable 3D information that can be used to aid in 3D face modeling. This project aims to demonstrate how the Kinect data can be used in conjunction with a generic face model to achieve an automated 3D face modeling process from 3D data. The project explored several methods of obtaining the data from the Kinect sensor as well as modifying the data of a 3D model by scripting. A framework was achieved whereby the user was able to generate a directly mapped and transformed 3D plane model using feature point data obtained from the sensor.Bachelor of Engineering (Computer Engineering
Computer-assisted auto coloring for 2D animation characters
No full textInking/coloring of the individual animated characters for each frame is one of the most time-consuming and labor-intensive procedures in cel animation production. Many systems and algorithms have been proposed to assist the coloring of cel animation production. However, these systems or algorithms are either not efficient enough, in terms of the degree of dependence on user intervention, or yield to significant limitation. Computer-Assisted Auto Coloring (CAAC) for cel animation remains a tough issue in research.DOCTOR OF PHILOSOPHY (SCE
- …
