In today's fast-paced society, many people are under tremendous pressure, and in the post-pandemic era, most are troubled by suboptimal health. Therefore, it is essential to offer people a viable lifestyle to relieve stress and enhance physical fitness.
Our Art Table Tennis is a product that transforms traditional table tennis into innovative interactive technology where users play table tennis games together but competitively to recreate art on a digital screen. This design aims to motivate people to exercise and enhance social connection between multiple players throughout the playful artistic process. The potential target users could be table tennis enthusiasts or general players who seek an enjoyable activity to release stress and exercise.
In the intended experience, a digital screen is mounted on the wall, divided into two canvas sections for two users to interact with. Each user hits balls towards their designated canvas. One user is required to wear a wristband equipped with a digital screen to initiate or restart the game. Upon starting the game, a 60-second countdown is displayed on the wall-mounted screen, signalling the users to begin their session.
The users stand by each side canvas must aim for their bottom areas to score highly (achieve 2 points). Hits outside these designated high-score zones on either canvas will result in a low score (achieve 1 points).
As the game progresses, points are visually represented on the screen, with each successful hit filling in segments of an art piece. Higher scores result in larger, sequentially revealed segments of the artwork from bottom to top. Conversely, lower scores reveal smaller segments. At the end of the 60-second countdown, the extent of the artwork's completion on each canvas is displayed. The user who has recreated the larger area of the painting is declared the winner.
Technical Description
The hardware of the entire device comprises a computer, camera, Arduino UNO, a 2.8-inch TFT Display Touch Screen, Arduino R4 WIFI UNO, multiple ultrasonic module sensors, and a projector. Additionally, for the physical setup, we have prepared multiple table tennis paddles and balls for multiplayer (doubles) gameplay. The electronic display screen, used to start and reset the game, is mounted on the user's wrist with a sports wristband.
To achieve the desired functionality, we have positioned ultrasonic sensors directly beneath the target whiteboard to detect any contact made by the balls. This data is transmitted to the computer via a serial port. Simultaneously, the camera connected to the computer captures and saves this frame, which is then processed by YoloV5S for image analysis. It checks the captured image against a pre-labelled image database for AI recognition to precisely determine and feedback on the exact impact location of the table tennis balls on the whiteboard. The main operational code for the AI recognition library is written in Python, utilising native Pytorch and CUDA Toolkit with Nvidia's virtual spaces to determine coordinates. Colour recognition of the table tennis balls is performed using OpenCV, identifying the colour and coordinates of the balls hit by one or two players in the frame. The process is optimised using Python’s Cudnn and TensorRT to ensure that the frame processing is completed within 40 milliseconds, maintaining the game's timeliness. After obtaining the colour and coordinates of the table tennis balls, the information is transmitted from the Python script acting as a server to Unity3D via sockets. The Unity 3D's C# script then responds to the game inputs, ultimately projecting the game's visual interaction onto the whiteboard through a projector, with sound effects delivered by the computer's built-in or external sound system. The Arduino UNO equipped with a 2.8-inch TFT Display touch screen Nike wristband directly communicates with Unity3D to send commands to start and replay the game.
By integrating various software and hardware technologies, we aim to deliver the best interactive effects and user experience to our players.