Akihiro Inui

Guitar sound syntheis using digital-waveguide model was computed in MATLAB. It consists of bidirectional delay line and the output is obtained by at any point in the delayline by summing the values.The physical string parameters such as string raduis, tension are given to emulate actual steel strings.

The automatic classification system to music genre and musical instrument was implemented with a set of low-level features and two supervised classification methods, k-Nearest Neighbors and Sparse Representation-based Classifier. It achieved 72 % for music genres and 99% for musical instruments as the highest recognition accuracy. Besides these approaches, another system with deep learning (Keras) was implemented in Python and it achieved 73% to music genre classification.

In 2017, "インスタ映え" (means photogenic on instagram) was nominated as the most popularly used word in Japan. In order to end the argument of people those who claim their photos are the best, building an automatic photogenic recognition system "Instafly" was executed.

It was featured on GIZMODO Japan

I am sometimes coding websites. Some examples I editted for my clients are below.