John He

Summary A creative and passionate graduate from UNSW who majored in Information Technology. Talented Junior Software Engineer adept at working with team members to accomplish trouble shooting based on client's report.

• Projects LSB based Steganography C++, OpenCV This project designed a program for users to hide one secret image in another typical graph. This algorithm only changed the lowest bit of the pixel’s value, making it 100% impossible for human beings to notice that a secret image is concealed in the normal one. After redesigning the algorithm and optimizing the code, the program’s running time is reduced by 50%. It can automatically resize the secret image when it’s too big for the standard image. Gesture Detecting based on Wi-Fi signal Python, Machine Learning, Tensorflow From a paper, I noticed that different gestures led to different regular signal loss around devices. Based on this feature, I designed several other gestures that all last 2 seconds and collected their signal loss using Wireshark. Then I pick 3 gestures that make the most special signal loss lines. A large number of data was collected to avoid coincidence, and some algorithms like mean filter, standard scaler was implemented to remove noise from the data. Neural networks was established using Tensorflow to learn and distinguish these 3 different signal loss lines. After adjusting the learning rate carefully, it reached a 70% accuracy, which was quite good compared with other gesture detecting methods. 1