M-3LAB / Real3D-AD

[NeurIPS 2023] Offical code for <Real3D-AD: A Dataset of Point Cloud Anomaly Detection>. A 3D point cloud anomaly detection dataset and benchmark.
107 stars 8 forks source link

How is the dataset used for multi-modal anomaly detection? #10

Closed zhaoyuan1209 closed 2 months ago

zhaoyuan1209 commented 2 months ago

Could you provide a colored point cloud that can be conveniently transformed into depth and RGB images? Alternatively, could you suggest a method to transform this dataset to meet the requirements for multi-modal anomaly detection?

shirowalker commented 2 months ago

zhaoyuan1209

It's not easy to provide this dataset extra color. If you use other tools to colorize, it is recommended to color the point cloud first and then map it to the 2D plane. But this is also not easy, and may even be more difficult than creating a synthetic RGB point cloud dataset.

zhaoyuan1209 commented 2 months ago

So, this dataset cannot be used to evaluate multi-modal anomaly detection tasks? (RGB+Depth image)

shirowalker commented 2 months ago

So, this dataset cannot be used to evaluate multi-modal anomaly detection tasks? (RGB+Depth image)

The dataset cannot be used to evaluate multi-modal anomaly detection tasks. The point clouds in the dataset have no colors.