Zhaoyibinn
Zhaoyibinn
Hello,I'm doing the similar things like you.  and i also faced "=> No good initial image pair found. => Relaxing the initialization constraints.". What's more,I have tried...
if not self.refine: return {"depth": depth, "photometric_confidence": photometric_confidence} else: resize_transform = torchvision.transforms.Resize(size=(depth.shape[1], depth.shape[2])) resized_img = resize_transform(imgs[0][0]) refined_depth = self.refine_network(resized_img.unsqueeze(0), depth.unsqueeze(0)) return {"depth": depth, "refined_depth": refined_depth, "photometric_confidence": photometric_confidence}
I have solved this problem, and the core issue is: The official PTH, TAR trained through official tutorials, and the structure of network inference. The names of the keys for...
> > 我已经解决了这个问题,核心问题是:我已经解决了这个问题,核心问题是: > > 官方PTH,通过官方教程训练的TAR,以及网络推理的结构。官方PTH,通过官方教程训练的TAR,以及网络推理的结构。三者的键名不同,读取时需要转换。 > > 官方的转换方法已经写在 Lightglue 库中,位于 Lightglue 的大约 470 行。py.官方的转换方法已经写在 Lightglue 库中,位于 Lightglue 的大约 470 行。py。 > > `for i in range(self.conf.n_layers):` ` pattern =...
> > 我现在也有和你一样的想法。但我更喜欢端到端方法来利用其实时性能。据我所知,目前只有 dust3r 支持 GT 姿势和内部函数。他们还有另一个工作 [pow3r](https://europe.naverlabs.com/pow3r.) 。也许这会对你有所帮助。 > > 嗨,从 dust3r 论文中,我认为他们不支持 GT 姿势,而且是固有的,因为它们使点标准化。我的理解正确吗? > > After carefully reading the alignment code of Dust3R, I found that it...
> > 祝贺这项出色的工作!泛化能力令人印象深刻,尤其是在户外场景下表现出色。 > > 关于我的数据集,我有地面实况内在函数和姿势,我的目标是获得具有绝对比例的深度图和点云。我有两个可能的方法: > > 是否可以将这两个参数(地面实况内在和姿态)合并到模型中,以直接获得具有绝对比例的深度和点云? > > 或者,我们可以将估计的姿势与地面实况姿势对齐以获得比例因子,然后将其应用于估计的深度,以实现绝对比例的深度图和点云。 > > 我非常感谢您对这些方法的建议。谢谢! > > 嗨,我们也在研究同样的问题,并且还使用了第二种方法。请问您得到了什么比例因子?我想知道它是否与所有样本一致,您也有这方面的发现吗? > > 非常感谢! I just used a similar approach to Dust3R to...
Oh, this is my question. Thank you very much for your attention. I have already committed the code with README and the corresponding two Colmap DTU datasets for your reference....
> > Oh, this is my question. Thank you very much for your attention. I have already committed the code with README and the corresponding two Colmap DTU datasets for...
> [@CanCanZeng](https://github.com/CanCanZeng) I plan to publish the 2D GS application project of AbsGS / MCMC, including other implementation details, on my GitHub, but not immediately. I will let you know...
Have you solved this problem?