NHL REBOUND PREDICTION
With goalies being so skilled in the NHL, one of the best ways to score a goal comes from a second chance on a rebound into a high danger area of the ice. Our team has developed a tool to accurately predict where a rebound will be directed after a shot. The tool will be used to train players to shoot in certain locations to produce the highest scoring chance following the shot.
The team trained a deep neural network, with help from TensorFlow libraries, using sample data consisting of shot location on ice, and shot location on net to then predict the rebound angle in 30 degree bins surrounding the net. The final product consists of a GUI which allows a user to choose where the puck is being shot from and the location on net, and output a spray chart of probabilities of where the resulting rebound will be.
View the paper associated with the project here!
RUBIK'S CUBE SOLVER
The overall purpose of the project was to implement a user-friendly interface that allows for the input of a 2x2 Rubik's cube and provides an intuitive step-by-step process for obtaining a solution to the cube. The application targets all users who are looking to solve a specific iteration of their personal 2x2 Rubik's cube without needing to learn a complex algorithm.
Within this project, I worked heavily on the computer vision aspect to be able to scan in each face of the cube and process the colours to be used by the solving algorithm. Other components of the project include the output frame displaying steps to solve, the title screen, and the manual input net.
Pictured below is the team's poster for the project.
FLAP-PY BIRD
Flap-Py Bird was one of the many games developed during the 2018-19 school year as a part of the QMIND Games Team. This addition to the collection of games was done by me with occasional input from others. The application gives the user the option to play the traditional Flappy Bird through PyGame code, or to watch a network of bots learn to play and essentially beat the game.
