whipbaek / movis-web

MOVIS Web

Capstone Design Project 2 : MOVIS (Mobility information Obtained by Very Intelligent System) for 2nd semester of 2022

• Role Sharing
  • Beak: Back-end & Server build
  • Hojun Song: Deep learning based car category classifier training & evaluate
  • Kim: Research classifier, Front-end & TTS
• Notion Record: (https://truth-scallop-0eb.notion.site/MOVIS-Mobility-information-Obtained-by-Very-Intelligent-System-df39c2449d394a1e96eeac24b7662f7f?pvs=4)

• Propulsion method
  1. Using car appearance data, public data shared on AI Hub
  2. Uploaded car images using VggNet, a representative image classification model, are classified by category and keyword
  3. Provide information in web crawls through labeled keywords in that image
  4. Show the information to the user and explain it in voice
• Expected effect
  1. It provides convenience by solving the procedure of searching in the browser after verifying the existing vehicle name with a single photo
  2. In the future, it is expected that this service, which provides a function to explain information about the vehicle in front of the driver through a black box or a camera installed on the vehicle, will be installed in the vehicle and fused
  3. It can also be used as a foundation technology in a system that classifies cars and does +alpha work