Simplify README and clarify frame extraction paths

README.md

@@ -50,15 +50,27 @@ bash scripts/extract_vlac2_release_raw_archives.sh /path/to/data
 ### 2. Extract frames
 
 ```bash
-python scripts/extract_vlac2_release_frames.py --data-root data
+python scripts/extract_vlac2_release_frames.py --data-root /path/to/data
 ```
 
-By default, extracted frames are written to `_extracted_frames/` under the release directory. The extractor materializes only the benchmark-designated main camera stream.
+This command reads raw videos from `/path/to/data` and writes extracted frames to `_extracted_frames/` under the release directory. The output location is fixed by the release package and does not need to be specified explicitly. Only the benchmark-designated main camera stream is materialized.
 
-If the raw videos were extracted elsewhere:
+For example, if the release directory is:
 
-```bash
-python scripts/extract_vlac2_release_frames.py --data-root /path/to/data
+```text
+/path/to/release
+```
+
+and the raw videos are unpacked to:
+
+```text
+/path/to/release/data
+```
+
+then the extracted frames will be written to:
+
+```text
+/path/to/release/_extracted_frames
 ```
 
 ### 3. Use the benchmark splits
@@ -91,34 +103,6 @@ should be resolved as:
 /path/to/release/_extracted_frames/...
 ```
 
-## Benchmark Splits
-
-The release contains one training split and four evaluation splits:
-
-* `train`
-* `test_expert_seen`
-* `test_expert_unseen`
-* `test_nonexpert_seen`
-* `test_nonexpert_unseen`
-
-The four test splits are intended for evaluation only.
-
-## Dependencies
-
-A video decoder backend is required for frame extraction. Install one of the following:
-
-```bash
-pip install opencv-python
-```
-
-```bash
-pip install av
-```
-
-```bash
-pip install imageio imageio-ffmpeg
-```
-
 ## Notes
 
 * This repository contains only the raw videos required by the released benchmark splits.