It is an open source project to provide opencv function on the web.
git clone https://github.com/JeHwanYoo/web-vision.git
This project was created using Nuxt.
# install dependencies
$ npm install
# migrate prisma (SQLite)
$ npx prisma migrate
# using prisma studio (SQLite)
$ npx prisma studio
# serve with hot reload at localhost:3000
$ npm run dev
# build for production and launch server
$ npm run build
$ npm run start
For detailed explanation on how things work, check out Nuxt.js docs and Prisma docs
Create a .env file on top of your project.
LIMIT_BODY_SIZE = 10mb
| enviroment | type | comment |
|---|---|---|
| LIMIT_BODY_SIZE | Memory Representation | After encoding the image as base64, you need to allow enough body size to be sent to the server. If the image size is larger than the allowed size, a 413 error will occur. |
This project requires opencv-python.
So, python and opencv-python must be properly installed on the server and it will use the server's resources.
If you are hosting online, consider the cost issue.
In 'opencv' directory, you can expand opencv function. Basically, I haven't implemented many features yet. More features will be added in future versions, and you can add your own.
If necessary, it would be a good idea to make the directory a Git submodule.
In 'opencv/api/client.ts' file, defines the interface of the opencv function to be executed on the client.
The same results are reused using caching.
The list is incremented only if the results are different.
Sometimes using a different pipeline can produce the same result. Because the results are the same.
- Basic Image Processing (0.1.x)
- Composite of Two or More Images (0.2.x)
- Histogram (0.3.x)
- Filtering (0.4.x)
- Finding Edges, Corner Points, ETC... (0.5.x)
- Video, Gif Extension Support (0.6.x)
- Webcam Support (0.7.x)
The plan can change at any time.
See CHANGELOG.md