DeepSense Forum
Hi and thanks for holding the competition,
I have joined the competition and have explored the dataset. I noticed that although we can download multi modal sensors' data, there is only the image information we can use for test.
Just want to make sure that this beam prediction competition is only related to the vision based dataset?
Many thanks,
Yao
Hi Yao,
Thank you for participating in the competition and getting in touch with us. The test data (which has still not been released) will have all the modalities and a similar data structure as the publicly available development dataset. As we mentioned on our competition page, this is a multi-modal beam prediction challenge. We do provide the flexibility to choose from the three different sensing modalities, i.e., vision, LiDAR, and radar, or to utilize any combination of these modalities to perform the said task.
We apologize if this wasn't clear from our problem formulation. Can you please point us to the section in our challenge description which gave you the impression that only visual data will be used during the testing phase? Your feedback would help us in addressing this concern for all participants.
Regards,
Team DeepSense
Sorry I was misunderstood by the vision-based example demonstration. I have no other questions.
Thanks for your kind reply.
Yao
Hi and what a great race and meaningful work!
Have the real beam results of the test data been released? Can you send me a copy?
My mail is quanzhou@bupt.edu.cn,
Many thanks,
Quan
