tusimple benchmark ground truth

[kivinju]

https://www.kaggle.com/datasets/manideep1108/tusimple?resource=download

The links below do not work anymore. Please use the Kaggle link above to download the dataset.

LANE DETECTION CHALLENGE

dataset

• train_set.zip
• test_set.zip
• test_baseline.json

ground truth

• test_label.json

VELOCITY ESTIMATION CHALLENGE

dataset

• benchmark_velocity_train.zip
• benchmark_velocity_test.zip
• benchmark_velocity_supp.zip
• baseline.json

ground trurh

• gt.json

[cardwing]

@kivinju , thanks for uploading the ground-truth labels.

[cardwing]

@kivinju , can you provide the evaluation codes of the lane detection challenge (scripts used in testing)? Or can you provide the details of using the test_label.json, e.g., how is a lane point classified as TP or FP? Thanks a lot!

[kivinju]

@kivinju , can you provide the evaluation codes of the lane detection challenge (scripts used in testing)? Or can you provide the details of using the test_label.json, e.g., how is a lane point classified as TP or FP? Thanks a lot!

@cardwing https://github.com/TuSimple/tusimple-benchmark/blob/master/evaluate/lane.py

[cardwing]

@kivinju , really appreciate your kind response.

[wcg5262]

@kivinju 您好，如果我想自己制作此类数据集，也要做您提供的这种json文件，我要从何入手呢？

[kivinju]

@kivinju 您好，如果我想自己制作此类数据集，也要做您提供的这种json文件，我要从何入手呢？

format your output and document.

[harryhan618]

Hi, thanks for sharing. May I ask what's the method used in baseline for lane detection?

[fabvio]

Hi, thanks for releasing the dataset. Just to let you know, I provide class labels for the lane boundaries in the training and validation set. Maybe someone is interested.

https://github.com/fabvio/TuSimple-lane-classes

[jiaofang1222]

@kivinju 您好，如果我想自己制作此类数据集，也要做您提供的这种json文件，我要从何入手呢？

您好，能分享一下生成.json文件的程序吗？

[obulirajs]

Hi, thanks for releasing the dataset. Just to let you know, I provide class labels for the lane boundaries in the training and validation set. Maybe someone is interested.

https://github.com/fabvio/TuSimple-lane-classes

As part of my academics i chose to do project on lane detection. I need to do labeling on the road lanes as part of my project work. i am pretty new to this domain and i do not know how to do. Can you help me in pointing right tool to do labeling and how to do. Any help is much appreciated.

-Thanks
Raj

[fabvio]

Well I just wrote my own labeling tool for this specific dataset using python :) if you need it please open an issue on my repo, side things like this shouldn't be discussed here.

[DGJSCUT]

Is there any possibility that the author could share the code or method of the dataset TuSimple about how to label it in a json format with us?

[Aya-S]

I would like to download only a small subset of the dataset any advice on how to do that?

[pandamax]

if i have lane instance segmentation images，is there a method to generate json format like Tusimple?thanks in advance.

[mohamadalii]

Hello, I would like to test Lane detection on real-time camera without the dataset, could anyone help me?

[eeyrw]

I wrote some simple code for walking through tusimple dataset and other popular datasets:
https://github.com/eeyrw/LaneDatasetInspector

[ntthuy11]

Hi TuSimple engineers, is it possible that you can share your annotation tool for lanes? I would like to train bike lanes by myself, and I am searching around for a good tool for lanes. I will follow TuSimple license. Thank you very much in advance!

[kivinju]

Hi, Thanks for asking. I am consulting our legal team for the license. Kai *Kai Zhou* Service Infra Manager ------------------------------ ------------------------------ *MWA* +1(619)916-3144 www.tusimple.com 9191 Towne Centre Dr, Ste 600, San Diego, CA 92122

…

On Thu, Mar 25, 2021 at 3:08 AM Siddharth Prajapati < ***@***.***> wrote: Hello I would like to know the license of the dataset? Is it possible to use TuSimple dataset for commercial purposes? Thank You — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#3 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABG5WPTHGDM5RVIEJWYIQODTFMDSDANCNFSM4F6HZ7KQ> .

[dexter2406]

@kivinju Isn't it Apache 2.0 license, as posted on the homepage of this repository?

[kivinju]

@dexter2406 Yes. I confirmed with our legal team and made it Apache 2.0.

[Frandre]

Thanks for sharing the dataset!

Could you please advice whether camera parameters, e.g. intrinsic, are available for your dataset or not?

Best,
Buyu

[wywywy01]

Thanks for your sharing.

It seems that in 'LANE DETECTION CHALLENGE', the data test_label.json line is 2782, while in test_set.zip only exists 2638 imgs.

could you please have a check?

[Takugo]

if i have lane instance segmentation images，is there a method to generate json format like Tusimple?thanks in advance.

Hi, have you found any tools or scripts that could generate TuSimple json format from lane instance segmentation images. I also have this sort of problems.

[miao02830]

如果我有车道实例分割图像，有没有办法生成像 Tusimple 这样的 json 格式？提前谢谢。

嗨，您是否找到了可以从车道实例分割图像生成 TuSimple json 格式的任何工具或脚本。我也有这样的问题。

同问，我也遇到了这样的问题，可以分享一下吗，十分感谢

[pandamax]

@miao02830 我这边实现了车道线实例分割数据集转tusimple的脚本，可能你的数据集标注形式类似于vil-100或curvelanes，可以参考一下vil-100转tusimple、curvelanes转tusimple

reference:

• convert vil-100 to tusimple
• convert CurveLanes to tusimple

[Takugo]

@miao02830 我这边实现了车道线实例分割数据集转tusimple的脚本，可能你的数据集标注形式类似于vil-100或curvelanes，可以参考一下vil-100转tusimple、curvelanes转tusimple

reference:

• convert vil-100 to tusimple
• convert CurveLanes to tusimple

@pandamax 嗯？环宇？

[harinduravin]

Hi, thanks for releasing the dataset. Just to let you know, I provide class labels for the lane boundaries in the training and validation set. Maybe someone is interested.

https://github.com/fabvio/TuSimple-lane-classes

Do you have labeled data for TuSimple Test set (test_label.json) as well? @fabvio
If possible, please share.

[zillur-av]

Does this dataset have lane class labels? Like dashed yellow = 1, sold yellow = 2, bot dotts = 3 or they are all considered as one class? I do not see any labels on the json files.

[josh-wende]

These links don't work for me, I get an "Access Denied" message. Is there somewhere else I can get the dataset?

[zehranrgi]

These links don't work for me, I get an "Access Denied" message. Is there somewhere else I can get the dataset?

https://www.kaggle.com/datasets/manideep1108/tusimple?resource=download

[SANGHEEMOON]

Hello,
I can't find 'test_baseline.json'. How can I get this file?

[02666]

How could i download these datasets?

[wy1663890176]

数据集下载不了，还有其他的方法获取数据吗

[lsj1012]

The dataset can't be downloaded, can you reply to it?

[kivinju]

Hi @lsj1012 @wy1663890176 @02666 @SANGHEEMOON I am no longer with TuSimple and it seems that the company has disable the links. Please use the Kaggle link to download the dataset: https://www.kaggle.com/datasets/manideep1108/tusimple?resource=download

[lsj1012]

嗨，我不再使用 TuSimple，似乎该公司已禁用链接。请使用 Kaggle 链接下载数据集：https://www.kaggle.com/datasets/manideep1108/tusimple?resource=download

Thank you very much

[wy1663890176]

你好，你下载下来了吗，我注册不了账号

嗨，我不再使用 TuSimple，似乎该公司已禁用链接。请使用 Kaggle 链接下载数据集：https://www.kaggle.com/datasets/manideep1108/tusimple?resource=download

谢谢

[wy1663890176]

Hi @lsj1012 @wy1663890176 @02666 @SANGHEEMOON I am no longer with TuSimple and it seems that the company has disable the links. Please use the Kaggle link to download the dataset: https://www.kaggle.com/datasets/manideep1108/tusimple?resource=download

谢谢

[lsj1012]

你好，你下载下来了吗，我注册不了账号

嗨，我不再使用 TuSimple，似乎该公司已禁用链接。请使用 Kaggle 链接下载数据集：https://www.kaggle.com/datasets/manideep1108/tusimple?resource=download

谢谢

The download is complete

[wy1663890176]

你好，你下载下来了吗，我注册不了账号

嗨，我不再使用 TuSimple，似乎该公司已禁用链接。请使用 Kaggle 链接下载数据集：https://www.kaggle.com/datasets/manideep1108/tusimple?resource=download

谢谢

下载完成

xiexie

[wy1663890176]

嗨，我不再使用 TuSimple，似乎该公司已禁用链接。请使用 Kaggle 链接下载数据集：https://www.kaggle.com/datasets/manideep1108/tusimple?resource=download

你好，想问一下，有生成的train_DRL_list.json和test_DRL_list.json这两个文件，我的环境运行会出错，生成的是空的，还有MyTuSimpleLane我文件夹里的所有文件都是空的

[02666]

The file of test_DRL_list.json and train_DRL_list.json is empty. How can I solve it