1
1

Delete article

Deleted articles cannot be recovered.

Draft of this article would be also deleted.

Are you sure you want to delete this article?

More than 3 years have passed since last update.

【Instance Segmentation】MaskRCNN : Instance Segmentationの基礎

Last updated at Posted at 2020-11-19

概要

![image.png](https://qiita-image-store.s3.ap-northeast-1.amazonaws.com/0/482094/ee80063d-6682-638e-8904-ad549cda123b.png)

Semantic Segmentationは知ってるけど、Instance Segmentationの仕組みがわからない。どうやって物体毎に分類するんじゃ?
てことでMaskRCNNをもう一度調べてみた。

使ったコードのリンク貼っておきます
https://github.com/yokosyun/instance-segmentation

semantic segmentaionについて知りたい方はこちら
https://qiita.com/minh33/items/6e42041dd5108d5fc2f0
https://qiita.com/minh33/items/d800394f480b2d6eef33

Instance Segmentation(MaskRCNNの場合)

![image.png](https://qiita-image-store.s3.ap-northeast-1.amazonaws.com/0/482094/fa6cf4e4-e5aa-6ebe-e3ee-285030995913.png)

RPNで推定されたBoundingBoxのエリアに対してのみsemantic segmentationする
=Instance Segmentation

めっちゃシンプル!

結果

personとしてsegmentationされているpixelがBoundingBoxの中にしか無いのが分かる。(実際はこのBBoxは補正された後のBBox)

image.png

結論

RPNで推定されたBBoxのエリア内にてSemantic Segmentationをする。

参考文献

Mask R-CNN https://arxiv.org/pdf/1703.06870.pdf https://github.com/Okery/PyTorch-Simple-MaskRCNN Mask R-CNN https://www.slideshare.net/windmdk/mask-rcnn
1
1
0

Register as a new user and use Qiita more conveniently

  1. You get articles that match your needs
  2. You can efficiently read back useful information
  3. You can use dark theme
What you can do with signing up
1
1

Delete article

Deleted articles cannot be recovered.

Draft of this article would be also deleted.

Are you sure you want to delete this article?