Xiang Chen

Python has become the most popular language for machine learning right now since almost all machine learning tools provide service for this language, and it is really to use since it has many build-in objects like Hashtable. In C, you need to implement everything by yourself.

C++ is one of the most popular programming languages in graphics. It has many fancy libraries like eigen to help us process matrix. I have many previous projects about graphics based on C++ and this time, we also need to deal with graphics since we need to analyze movements of the human body. C++ has much more advantages than Java. C++ uses only compiler, whereas Java uses compiler and interpreter in both. C++ supports both operator overloading and method overloading whereas Java only supports method overloading. C++ supports manual object management with the help of new and delete keywords whereas Java has built-in automatic garbage collection.

OpenCV is a library of programming functions mainly aimed at real-time computer vision. It is really famous in the industry realm nowadays. Many high-tech companies have chosen it as the primary developing tool, so do I.

I have learnt related courses in computer vision and graphics and OpenCV is the key to the success of these courses. It can provide us with highly efficient noise filters, convolution matrix and classical models. In my current project, I need to analysis the movement of the human body, which will include noise filtering, object recognition, and skeleton construction. OpenCV can provide us with HOG and other build-in models to make my life easier.

Google Cloud Platform has comprehensive features. It can provide us with static IP, Computer Engine, and GPU for this project. It also gives us 300 $ for free to use.

I personally have previous developing experience with the Google Cloud Platform. I found it is really convenient and all features have clear documentations. I used the computer engine feature to build a virtual machine for my project. The virtual machine is really stable and secure. During this time search, I also find GCP has powerful data analytics ability, and even more, it has a well-constructed ecosystem.

Pytorch is a famous tool in the realm of machine learning and it has already set up its own ecosystem. Tutorial documentation is really detailed on the official website. It can help us to create our deep learning model and allowed us to use GPU as the hardware support.

I have plenty of projects based on Pytorch and I am familiar with building deep learning models with this tool. I have used TensorFlow too but it is not dynamic. Tensorflow works on a static graph concept that means the user first has to define the computation graph of the model and then run the ML model, whereas PyTorch believes in a dynamic graph that allows defining/manipulating the graph on the go. PyTorch offers an advantage with its dynamic nature of creating graphs.