1. Introduction

1.1 Overview

Nowadays machine learning across a range of tasks is very powerful and can serve our life. 2 years ago, our information-gathering behavior had been to find out what other people think through "Opinion Mining and Sentiment Analysis". So we have a hypothesis about whether we can figure out the the meaning of an image. Tensor-Flow is such an open-source software library for us to detect and decipher patterns and correlations. Thus, the module which was trained based on Big Data should be accurate and the data must be rich. In our life, we always type a keyword in Google and Google will show the photos about it. However, if we have a photo and we don't understand what is that, it's hard for us to figure it out. Therefore, one of the example by using Tensor-Flow is to classify the image. Also, you can use Tensor-Flow to classify voices and signals. Tensor-Flow can be written in several programming languages such as Python, C++, Java, GO and so on. By using Tensor-Flow, our life will become more and more convenience. Based on that, we can take use of Tensor-Flow to find out what an image's meaning is. Thus, our project aims to analyze and recognize an amount of user' profile photo, which is provided by Weibo API from MPI UB Lab. Based on these result, we may know some preference of the users[1], such as what clothes the user dressed, what jewelry the user wear. After that, we can offer the result to Weibo so that Weibo are able to provide the related advertisement according to user's preference.

1.2 Tools

Programming language: Python, SQL
Database: MySQL
Server: Apache HTTP Server