Identifying and Classifying School Uniforms
About the Project
Human trafficking is a worldwide problem. Even today, there are upwards of 45 million people that are currently enslaved. In the United Kingdom, the Home Office (a government department responsible for immigration, security, and law and order) has a database of over 17 million images and videos of child abuse material. Of that 17 million, a significant number of images contains children in their school uniforms, which can be used as identifiers for the school that child belongs to. The overall goal of this project is to develop a model pipeline that identifies if a child is wearing a uniform, and if so, identify what school that uniform belongs to so action can be taken.
There are many teams working on different stages of this pipeline. For our contribution, we have been focusing on detecting the location of different articles of clothing in the image and identifying features such as color or pattern. Since most schools in the United Kingdom have a specific dress code, we can narrow down the possible schools a uniform belongs to from these features.
Figure 1 comes from a clothing detection API developed by Clarifai, an artificial intelligence company specializing in computer vision. Using their API, we can upload an image and get bounding boxes on detected pieces of clothing, which we can further analyze.
About the Sponsor
The Global Emancipation Network fights human trafficking – a true modern evil. Approximately 21 million men, women, and children are trafficked each year across the globe generating an estimated $50 billion for traffickers and organized criminal networks. The International Labour Organization (ILO) estimates that internationally, fewer than 50,000 of these victims are identified and rescued. Like many other criminal enterprises, traffickers are reliant on internet-based applications to find and groom victims, organize transportation, and advertise their victims’ services. This reliance is also their weakness.
At Global Emancipation Network, we strive to be the global clearinghouse for trafficking data worldwide, as well as to facilitate communication and technology-sharing initiatives between the numerous anti-trafficking stakeholders across the globe.