GAST-Net-3DPoseEstimation
GAST-Net-3DPoseEstimation copied to clipboard
fabro66
A Graph Attention Spatio-temporal Convolutional Networks for 3D Human Pose Estimation in Video (GAST-Net)
gast5/lib/pose/hrnet/pose_estimation/gen_kpts.py", line 342, in generate_ntu_kpts_json pose = preds[num].tolist() IndexError: index 1 is out of bounds for axis 0 with size 1 How do I solve this situation where I want...
hi, thank for ur brilliant repo! when i draw 3d prediction image refer ur render_animation() function, like this def render3d(pose): fig = plt.figure() ax_in = fig.add_subplot(111, projection='3d') ax_in.view_init(elev=15., azim=70) ax_in.get_xaxis().set_visible(False)...
about R=rot, t=0 (rot = np.array([0.14070565, -0.15007018, -0.7552408, 0.62232804], dtype=np.float32)) R and T are parameters that determine the camera Does this mean that when I create a dataset, the position...
Is there any tests done in real time scenario? How much is the fps in real time scenario?
I ran your code quite differently from the paper,so I think there is something wrong with the data processing part. The data_3d_humaneva15.npz and data_2d_humaneva15_gt.npz files I obtained by VideoPose are...
https://github.com/fabro66/GAST-Net-3DPoseEstimation/blob/master/INFERENCE_EN.md The following problems occurred after the configuration was run Traceback (most recent call last): File "gen_skes.py", line 180, in generate_skeletons(video=video_path, output_animation=args.animation, num_person=args.num_person) File "gen_skes.py", line 155, in generate_skeletons viz_output,...
https://github.com/fabro66/GAST-Net-3DPoseEstimation/blob/946f6b701452204d23969a43dae348b69eca9bd9/main.py#L225 Dear author, thank you for your amazing work! I feel confused about the code in the train fuction.You seem to set the first 3D keypoint in groundtruth as (0,0,0).What...
@fabro66 I want to use these 3D coordinates to manipulate 3D structure in unreal engine. please help me to understand the nature and unit of these coordinates? are those normalized?
Hi, It is a great work, and here I used the mediaPipe to instead the Yolov3+HrNet for testing the model in the wild video, the effect is pretty good with...
