Sicarius
Sicarius is a fast-paced, top-down shooting game. Set in the post-apocalyptic world, you have control over a spider to defeat the killer robots and retrieve the Tree of Life.
Faculty of Engineering
Quantum Education Society @ 9th Inno Show
We seeks to transform quantum physics education through the creation of an innovative curriculum that prioritizes experiential learning and practical engagement. The Innow Show offers a unique platform that aligns perfectly with our project’s goals of enhancing quantum physics education through experiential learning. By incorporating practical experiences and demonstration kits, our goal is to empower students with a profound understanding of quantum technology, effectively preparing them for future opportunities in this burgeoning field.
VR Motion Capture System For Third-Person Perspectives
A VR Motion Capture System For Third-Person Perspectives, aims to analyze performance in term of VR sickness incidence, Avatar embodiment and User Engagement
of Embodied Users in CAVE System
Talk Your Way Out: A Puzzle Game Demo with LLM NPC
Talk Your Way Out, a game demo demonstrating the capabilities of AI in gameplay and storytelling. Talk with non-player characters (NPCs) to solve puzzles and find a way to escape. The game uses large language models, speech recognition and text-to-speech to provide dynamic, natural, and quick conversations with the NPCs, demonstrating capabilities of AI in simulated environments.
You are an alien named Blorb Gorb, you woke up finding yourself captured by scientists of the Earth in a laboratory somewhere in Hong Kong. You have to escape, not just using brute forces, but also your disguise ability…
Transknight
We implemented a 2D platformer game with creative character controls, high quality sets of enemies and maps design, and a complete game play process that can give players satisfaction and sense of achievement when they defeat enemies, solve puzzles, and clear the game. Our main character is a druid-like knight that can transform into 3 other different animals, each with unique skillsets and playing style. Across 5 different levels, players are going to encounter varied types of enemies, defeating them and unlocking new power. After defeating the final boss and clearing all the levels, you will become the savior of the land and the true hero for the people.
Breathing in Virtual World
This project looks into VR’s potential as a Chinese language learning tool for students whose first language is other than Chinese. Particularly, the study focuses on the efficiency of VR based educational games to raise Hong Kong’s non-Chinese local secondary school students’ confidence in interacting in Chinese and using it for common daily tasks. The project is done in collaboration with IBEL – a center that provides educational support to underprivileged non-Chinese Speaking (NCS) communities in Hong Kong.
Scavenger 01_Automatic ping pong collector
SCAVENGER-01 uses AI and computer vision to perform object detection and pick up ping pong balls, without interrupting athletes’ training.
It saves time, improves training efficiency, and prevents injuries.
It can be deployed in ping pong arenas, ideal for ping pong training centres, schools, colleges, clubs, and home settings. Where the robot will pick up the ping pong balls while athletes can focus on their training.
SCAVENGER-01 is a prototype, in the future, we plan to deploy a human avoidance algorithm to avoid colliding into the path of athletes, improve the routing algorithm so it can effectively go through the entire map/arena, and automatic wireless charging so the unmanned robot can head to the wireless charging station for automatic recharging.
Smart PV Panel Cleaning Robot
Solar farms require flat and open spaces, however, Hong Kong is a relatively mountainous region. With Hong Kong’s geographical structure, it is not feasible for Hong Kong to develop hugely in solar farms. Nevertheless, we can expect a rise in the use of solar PV panels on village house rooftops and building rooftops by individuals for having partially self -sufficient energy systems to reduce electricity charges. Therefore, we have a great business opportunity here. Supporting the maintenance of solar PV panels with a cleaning robot is a meaningful idea at the right time. Moreover, doing manual cleaning for solar PV panels on rooftops is a slow and risky process. People’s safety is a major concern as working from heights and in different weather conditions can hinder people’s safety when cleaning the solar panels. Moreover, cleaning done by people is hard and slow process. With a cleaning robot, the cleaning work becomes simple and a fast process.
With a great number of policies and measures being implemented to encourage the use of renewable energy, it is a time where we could expect more installations of solar PV systems. This would be a fantastic opportunity for our team to make a solar PV panel cleaner to enhance the maintenance of the solar PV panels. With our product, it brings convenience to 6 people having solar PV systems on their home’s rooftops or buildings. This could further encourage people to opt for renewable energy from solar PV panels.
Machine Learning in Archaeology
Digitization of archaeology is in great demand. Since 2009, an HKU archaeological team of researchers and students led by Dr. Cobb has been investigating the area around Vedi, Armenia, aiming at understanding human life and mobility in the ancient landscapes of the Near East. A large volume of sherds was excavated and documented with photography. Inspired by the recent advancement in computer vision and machine learning, this project attempts to explore various deep learning models to classify and compare those sherds unearthed. It is hoped that insights gained from the project can help archaeologists of manage the massive quantity of ancient artifacts in the future.
SLAM robot with monocular depth estimation
This project aim to construct a robot using SLAM and a monocular camera for both SLAM and depth estimates. The system will be able to map the environment correctly and robustly, determining the robot’s location and orientation inside it. The robot can then estimate the depth using a deep learning model and pictures recorded by the monocular camera as input. This research intends to develop robotics and computer vision by investigating the potential of monocular cameras for SLAM and depth estimation, which might have practical applications in autonomous navigation and depth estimation through the use of a neural network model.