Image-Text-Nontext-Classifier-Service

Text/non-text image crawler & classifier service. Software course project of EIC, HUST (2021 Autumn).

Features

Neural Network Classifier

See https://github.com/KSkun/Image-Text-Nontext-Classifier.

Search Engine Spider

See https://github.com/KSkun/Image-Spider.

Web User Interface

Frontend: https://github.com/KSkun/Image-Classifier-Frontend.

Backend: https://github.com/KSkun/Image-Classifier-Backend.

Docker Containerize

The service is split to 4 modules and developed individually, then containerized with Dockerfile. In this repo, the service is configurated with docker-compose.yml, which makes deployment easier for users.

Configuration

1. Install git and git-lfs.
2. Pull this repo with command git clone --recurse-submodules.
3. In ./classifier folder, run git lfs pull.
4. Check if you need a third-party source for pip or npm in Dockerfile. Note that frontend/Dockerfile uses a HTTP proxy server by default, remember to configurate it for your environment.
5. Set the port you want in ports field of frontend in docker-compose.yml.
6. Run docker-compose up -d, wait for building image & creating containers. If you need a development environment, use docker-compose-debug.yml instead.
7. Test it with your browser.