We are conducting research on speeding up image processing. Image processing is now used in every situations such as artificial intelligence, autonomous driving, IoT, video conferencing, and surveillance cameras. The more advanced image processing, the greater the amount of calculation, and higher speed is essential. In addition, automatic driving needs to be judged before an accident occurs, and even simple image processing needs to be calculated at extremely high speed. In addition, devices such as IoT do not contain large programs, and in addition, it is necessary to save power, so it is necessary to change to an algorithm that runs at high speed or change the program. In our laboratory, we are researching algorithms to automatically speed up image processing designed according to the purpose, and image processing to automatically convert image processing programs so that they can be calculated on various platforms. We are conducting research and development on dedicated compilers and dedicated programming languages.