how to visualize the results as the demo.gif ?

[GuangtaoLyu]

hello, thank you for your great work. i run the render_blender.py and i get the result as follows. how can i get the result like the demo or the picture in the paper?

[y-zheng18]

what does your result look like? Seems you attached two images from the paper and demo

[GuangtaoLyu]

I run the eval.py. i got some obj and ply files. i also try the vis script in README. i got the many images.  but no color and scene？

---Original--- From: @.> Date: Tue, May 21, 2024 00:33 AM To: @.>; Cc: "Guangtao Lyu ( 吕光涛 @.**@.>; Subject: Re: [y-zheng18/GIMO] how to visualize the results as the demo.gif ?(Issue #7)

what does your result look like? Seems you attached two images from the paper and demo

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>

[y-zheng18]

you might need to use the original scene mesh to render the results you want, not the point cloud. For example, you can use ./data/GIMO/classroom0219/scene_obj/textured_output.obj which has color information.

[GuangtaoLyu]

thank you. i will try it.

---Original--- From: @.> Date: Tue, May 21, 2024 00:42 AM To: @.>; Cc: "Guangtao Lyu ( 吕光涛 @.**@.>; Subject: Re: [y-zheng18/GIMO] how to visualize the results as the demo.gif ?(Issue #7)

you might need to use the original scene mesh to render the results you want, not the point cloud. For example, you can use ./data/GIMO/classroom0219/scene_obj/textured_output.obj which has color information.

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>

[GuangtaoLyu]

you might need to use the original scene mesh to render the results you want, not the point cloud. For example, you can use ./data/GIMO/classroom0219/scene_obj/textured_output.obj which has color information.

how to visualize the results like this.

[GuangtaoLyu]

hello, zheng yang:

                                                                               fig 1

as the fig.1 in your paper, you say that you collect a diverse range of daily activities.  i don't find the text description in the dataset and i only find some words in the dataset.csv same as your reply in the issue#3.

Does each motion sequence correspond to only one activity in the fig. 2?

In our experiments, we predict the future motion in 5 seconds from 3 seconds input, where the first 3 seconds of a trajectory is just about to start an activity (i.e., beginning to move for fetching a book) in our dataset, and in the next 5 seconds the trajectory proceeds to finish the activity. We set the motion frame rate to 2 fps, i.e., 6 pose input and 10 pose output. Note that once the waypoints are predicted, a full motion sequence with high fps can be easily generated [51]

              fig 2

Message ID: @.***>

[y-zheng18]

Each motion sequence corresponds to one activity defined in Tab.2. Some sequences have text descriptions in dataset.csv.

[y-zheng18]

you might need to use the original scene mesh to render the results you want, not the point cloud. For example, you can use ./data/GIMO/classroom0219/scene_obj/textured_output.obj which has color information.

how to visualize the results like this.

For this, project the 3D gaze point to images to visualize 2D gaze.