0
0

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.

【Stereo Depth】左右の特徴量を結合してUnetでDisparityを推定させてみた

Last updated at Posted at 2020-12-09

背景

Stereo Depthで4D Cost Volume(Disparity,Channel,Height,Width)を作る手法が主流となっているがそれだと遅いんじゃないと思い今回自分で考えたネットワークを学習させてみた。

シンプルに右と左の特徴量をConcatenateしてUnetでDisparityを求めるだけのシンプルなネットワークになっている。

結果

KITTI Datasetを100 epoch Self-SuperviseでTrainingした。Batchはメモリーの都合上1に設定。

run time = 20.9271907806[ms]

input left

![left.png](https://qiita-image-store.s3.ap-northeast-1.amazonaws.com/0/482094/afe7dfab-84bb-f487-4376-c9cd33751333.png)

input right

![right.png](https://qiita-image-store.s3.ap-northeast-1.amazonaws.com/0/482094/5e06346f-3483-053f-83bc-bd30894c556a.png)

disp left

![disp_left.png](https://qiita-image-store.s3.ap-northeast-1.amazonaws.com/0/482094/bc9daf49-0f84-636c-f871-50b2f97ed2b0.png)

disp right

![disp_right.png](https://qiita-image-store.s3.ap-northeast-1.amazonaws.com/0/482094/5c4fedf5-aec0-4285-656d-16ee841d94a3.png)

Reconstruct Left

![estLeft.png](https://qiita-image-store.s3.ap-northeast-1.amazonaws.com/0/482094/84ffcec4-601d-f64e-6115-044467929941.png)

Reconstruct right

![estRight.png](https://qiita-image-store.s3.ap-northeast-1.amazonaws.com/0/482094/f2d6aeb1-85cc-4afe-ba77-0c92f241d850.png)

結論

・あんまり結果が良くないですね まだまだ改善の余地ありそうです
0
0
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
0
0

Delete article

Deleted articles cannot be recovered.

Draft of this article would be also deleted.

Are you sure you want to delete this article?