[PDF]

A Cross-platform Face Analysis Sticker App


Zhenghong Ye

07/10/2021

Supervised by Alia I Abdelmoty; Moderated by Richard Booth

In the pandemic, we can add a mask to our avatar to kindly remind people to wear a mask in order to be united to fight the virus together and bring social benefits. At Christmas time, we can send our friends and family an avatar with a Christmas hat to convey happiness. However, the current applications for adding masks to avatars has some limitations. For example: 1. Recognition errors result in masks being added in the wrong place. 2. Masks cannot be adjusted once added. 3. The choice of stickers is very limited. 4. No share feature once created. 5. Complex operation and time-consuming.

Therefore the goal of this project is to build a cross-platform avatar making app with possible solutions to tackle the issues above. It can also be shared with friends and family directly. Cross-platform and no downloading required functions based on Taro solution, it will be adapted to WeChat mini programme and web app. Back-end storage service and AI face recognition based on Wechat mini programme cloud and Tencent intelligent services.

Featured functions: 1. Auto-wearing mask once photo uploaded. 2. AI face recognition, features analysis. 3. Create poster pages that allow users to share and support opening on the web. 4. Different types of stickers are provided.

Timeline: Stage0: Literature review Stage1: User personas Stage2: UI/UX Stage3: Project configuration Stage4: Development Stage5: Report

More information for this project: Mini program: https://www.abacusnews.com/who-what/mini-programs-apps-inside-apps-make-wechat-so-powerful/article/3000920 https://www.dragonsocial.net/blog/wechat-mini-programs/

Tencent AI platform: https://ai.qq.com/hr/ailab.shtml


Final Report (07/10/2021) [Zip Archive]

Publication Form