[IS-Net] Contours that stand out in every corner, the latest version of the cutout segmentation model [Inference Colab, with CoreML model]
Cut out the objects in the image neatly
There is a model that cuts out prominent objects in the image neatly and segments it just by inputting the image .
It can be used for various purposes such as image composition if it can be separated cleanly to every corner like die cutting.
Latest high-precision model IS-Net
In the past, there was a fairly accurate semantic segmentation model called U2-net, which was used in a variety of applications.
A high-precision model by the same author was announced in July 2022.
It is IS-Net.
(Quoted from the official repository)
How to use
Python
Reasoning Colab
Run the following Colab notebook and the results will be saved in the results folder.
Alternatively, you can clone the repository, download the pre-trained model, and then infer with the script below.
Swift
Models converted to CoreML can be used on iOS and MacOS.
Make an inference request.
🐣
I’m a freelance engineer.
Work consultation
Please feel free to contact us with a brief development description.
rockyshikoku@gmail.com
I am making an app that uses Core ML and ARKit.
We send machine learning / AR related information.