ROMP
ROMP copied to clipboard
Arthur151
Monocular, One-stage, Regression of Multiple 3D People and their 3D positions & trajectories in camera & global coordinates. ROMP[ICCV21], BEV[CVPR22], TRACE[CVPR2023]
Hello, I have several questions regarding the pretraining phase. I would like to use pre-trained `hrnet` or `resnet50` models. 1. the `pretrain_hrnet.pkl` mentioned in many previous issues is not in...
In image_base.py, when you calculate occluded person, I think there is wrong condition. full_kp2ds is normalized to [-1 ~ +1] before passed to detect_occluded_person() as belows: ```python def process_kp2ds_bboxes(self, full_kps,...
Hi @Arthur151 What is the meaning of following error which I encountered in google Colab? I think the yaml file is missing in the active_config directory. Please check the issue....
question about the difference between the intrinsics and extrinsics calibrated and estimated by ROMP
Hello, I use ROMP to estimate the SMPL and use it to render novel view pictures. But I found that if using the intrinsics and extrinsics estimated by ROMP can...
蒸馏模型
我们想通过蒸馏的方式训练一个模型，看到您做过相关的工作： 还有一种简洁高效的思路，就是模型蒸馏，直接蒸大模型的输出就好了，我试过蒸小模型，很有效，很快就能训好。 请问您用的哪个模型作为teacher模型以及蒸馏相关的花费方程如何设置会比较好？
您好，我在翻一些数据集(比如coco)的数据处理的时候发现您好像没有传入相机内参。这让我有点疑惑，想确认一下这是否是我看错了。 如果您在预测时确实没用内参，那么我想询问一下就是照片上相同大小的人，在不同内参下，深度是不一样的。但BEV似乎是可以直接预测深度的，请问您是怎么解决这件事情的？
In function ' coarse2fine_localization', for adjusting z-wise only, I think we use cam_maps_3d[:, :, :, :, 0], not [:, :, :, :, 2] because self.coordmap_3d is the result of concatenate...
您好，运行`romp --mode=video --calc_smpl -i=path/to/input -o=path/to/output`，使用的是ROMPv1模型，但是我想要更换为ROMP_HRNet32_V1.pkl模型，请问从哪里下载？又在哪里更改代码呢？
大佬好，我在阅读论文中有些困惑，我对于3DPW评估中的三种协议是这样理解的 协议一：不用3dpw-train数据进行训练，对整个3dpw数据集进行评估 协议二：不用3dpw-train数据进行训练，对3dpw-test进行评估 协议三：使用3dpw-train数据进行训练，对3dpw-test进行评估 请问是这样吗？
