Clarification on Cluster Grouping and Map Status (Before/After Change)

[ShenZheng2000]

Hi, thanks again for sharing the dataset. After reading the paper, I had two questions:

(1) For each cluster in scene_clusters.py, how are the run segments grouped? Is it based on location only, or also on map change type? (I didn’t find map change types for each cluster).

(2) For each run segment, I don’t know whether its map is outdated, since I don’t know when the map change happened. Is there a way to tell if a run is before or after the change?

[johnwlambert]

Hello @ShenZheng2000, re (1), the scene clusters are grouped purely by location only. One location may have one or more change types.

We discuss this a little more in the FAQ.

Q: Is there a file that indicates different log pairs and what has changed?

A: We provide a clustering of logs by spatial location here. A few things to note:

Each log within a cluster shares some significant visual overlap with other logs within its cluster.

These are not necessarily before/after pairs. In some cases, all logs in a cluster may be "after" a change.

Each cluster has at least one log in the val or test set.

Logs of each cluster are provided in chronological order.

[ShenZheng2000]

Thanks. A follow-up to question (2): For each run segment, do you provide the map before or after the map change? My understanding is that both the outdated map and the updated ground-truth map should be available, but currently, each folder seems to contain only one map.