Skip to content

[Opencv.js doc] Init commit to add semantic segmentation example#359

Merged
huningxin merged 1 commit intohuningxin:gsoc_2020_dnnfrom
akineeic:gsoc_2020_dnn
Jul 25, 2020
Merged

[Opencv.js doc] Init commit to add semantic segmentation example#359
huningxin merged 1 commit intohuningxin:gsoc_2020_dnnfrom
akineeic:gsoc_2020_dnn

Conversation

@akineeic
Copy link
Copy Markdown
Collaborator

Fix #316

@akineeic
Copy link
Copy Markdown
Collaborator Author

I have tested the models for semantic segmentation in the wiki. The ResNet101_DUC_HDC can't be read correctly. The output image of Enet doesn't look good visually, and the inference time of FCN is too long (need about six minutes). So I just use the deeplabV3 model you mentioned here for the js example. The results are shown below.

  • Enet

Enet

  • FCN

FCN

  • DeeplabV3

DeeplabV3

@huningxin
Copy link
Copy Markdown
Owner

So let's use the deeplab v3 as the default model. BTW, did you see similar results in native python example? Please update that in #315 . And please report corresponding issues to upstream opencv for more guidelines.

@akineeic
Copy link
Copy Markdown
Collaborator Author

Thanks for your reminder. I update the results and put the upstream issue into #315.

@huningxin
Copy link
Copy Markdown
Owner

That's good. Let's merge this PR.

@huningxin huningxin merged commit 7892cc0 into huningxin:gsoc_2020_dnn Jul 25, 2020
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants