Download

.gitattributes

@@ -62,4 +62,3 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.rad filter=lfs diff=lfs merge=lfs -text
 *.ply filter=lfs diff=lfs merge=lfs -text
 *.rrhd filter=lfs diff=lfs merge=lfs -text
-*.fbx filter=lfs diff=lfs merge=lfs -text

README.md

@@ -38,15 +38,6 @@ tags:
 
 # Hard Intersection Multimodal Samples
 
-
-## Release Notes
-|Release|Description|
-|-|-|
-|v1.0.0|Initial public release.|
-|v1.1.0|Added Unreal Engine assets.<br>Fixed issues in the OpenDRIVE map data.<br>Updated the README to improve documentation and usability.|
-|v1.2.0|Added OpenCRG road surface elevation data.<br>Updated 3DGS reconstruction assets.<br>Updated the README to document OpenCRG support.<br>|
-
-
 ## Dataset Summary
 Hard Intersection Multimodal Samples is a curated multimodal dataset of accident-prone urban intersection in Japan for autonomous driving research and development.  
 It provides multi-camera images, trajectory, HD maps, semantic annotations, point cloud data, and 3DGS assets.
@@ -74,7 +65,7 @@ The current release include the following intersection:
   - Six synchronized camera views  
   - Two synchronized camera views
   - Trajectory data  
-  - HD maps (OpenDRIVE+OpenCRG/RoadRunner/Lanelet2/Vissim/Unreal Engine)
+  - HD maps (OpenDRIVE/RoadRunner/Lanelet2/Vissim)
   - Semantic annotations  
   - point cloud data 
   - 3DGS reconstruction assets  
@@ -89,7 +80,7 @@ The current release include the following intersection:
   - semantic projection domain  
   - reconstruction domain  
   - 360-degree coverage with six cameras  
-  - Developer-friendly formats: OpenDRIVE+OpenCRG/RoadRunner/Lanelet2/Vissim/Unreal Engine
+  - Developer-friendly formats: OpenDRIVE/RoadRunner/Lanelet2/Vissim  
   - LAS point clouds  
 
 ---
@@ -102,7 +93,7 @@ The current release include the following intersection:
 - Trajectory data
 
 #### Derived assets
-- HD maps (OpenDRIVE+OpenCRG/RoadRunner/Lanelet2/Vissim/Unreal Engine)
+- HD maps (OpenDRIVE/RoadRunner/Lanelet2/Vissim)
 - Semantic image labels
 - Semantic point clouds with labels
 - 3DGS reconstruction assets  
@@ -133,24 +124,11 @@ The current release include the following intersection:
 ---
 ### HD Map Formats
 ![hdmap](assets/lanelet2.png)
-
 The HD map is provided in the format shown below.
 - **OpenDRIVE(Ver1.4/1.6/1.8)**  
-- **OpenDRIVE(Ver1.4/1.6/1.8) + OpenCRG**  
 - **RoadRunner HD Map**  
 - **Lanelet2**  
 - **Vissim**  
-- **Unreal Engine**  
-
----
-
-### OpenCRG
-
-OpenCRG road surface elevation data is provided. 
-
-The data contains road surface height information (z(u,v)) derived from high-precision MMS measurements and can be used together with OpenDRIVE.
-
-![opencrg](assets/opencrg.png)
 
 ---
 
@@ -223,9 +201,6 @@ The data contains road surface height information (z(u,v)) derived from high-pre
 - System Scale: Scaled beyond single objects to cover large-scale road networks through an integrated image and point cloud pipeline. Furthermore, when paired with our HDMap data, it functions as a "Metric 3DGS" capable of precise real-world physical dimensions and positioning.
 
 ![3dgs](assets/3dgs.png)
-
-By generating 3DGS assets, it is possible to create videos or images by providing trajectories, such as a [straight path](https://youtu.be/wLVq0r28MmQ) or a [right-turn path](https://youtu.be/ZNdhgmF8uLI).
-
 [Please refer to the following viewer for 3DGS assets.](https://huggingface.co/spaces/dynamic-maps/hard-intersection-3dgs-sample)
 
 ---
@@ -240,9 +215,7 @@ Dataset
 ├─maps
 │  ├─lanelet2
 │  ├─opendrive
-│  │   └─opencrg
 │  ├─roadrunnerhdmap
-│  ├─unrealengine
 │  └─vissim
 │      └─images
 ├─pointcloud
@@ -307,6 +280,11 @@ This dataset was generated using several open-source models and libraries. We gr
   License: Apache License 2.0  
   https://github.com/IDEA-Research/GroundingDINO
 
+- **OneFormer**  
+  Jain et al., SHI-Labs, CVPR 2023  
+  License: MIT License  
+  https://github.com/SHI-Labs/OneFormer
+
 - **ViTMatte**  
   Yao et al., HUST Vision Lab, Information Fusion 2024  
   License: Apache License 2.0  
@@ -318,6 +296,11 @@ This dataset was generated using several open-source models and libraries. We gr
   License: Apache License 2.0  
   https://huggingface.co/IDEA-Research/grounding-dino-base
 
+- **OneFormer Cityscapes Swin-L**  
+  SHI-Labs  
+  License: MIT License  
+  https://huggingface.co/shi-labs/oneformer_cityscapes_swin_large
+
 - **ViTMatte Base (Distinctions-646)**  
   HUST Vision Lab  
   License: Apache License 2.0  

assets/jp_tokyo_takanawadai-lod.rad

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0eeae3dcdb391219dff3be802aa11c9912d33263ca81fa5f68e877add57ed729
-size 1185784000
+oid sha256:b2efbfcf21f38818953c5f9be6d84036d5ce621aaa8c4fe7429b3b93d0e6595e
+size 170618896