Match the blog copy and order

README.md

@@ -24,16 +24,32 @@ configs:
 
 # EgoHome: 2 hours of human-annotated egocentric video
 
-![The four annotation types on the same 5 s of one clip: hand boxes, subtasks, hand actions and hand captions](assets/egohome_header_2x2.gif)
+![The four annotation types on the same 5 s of one clip: hand boxes, subtasks, hand actions (Ego4D verbs), hand captions.](assets/egohome_header_2x2.gif)
 
-EgoHome is 1.99 hours of first-person household chore video (137 clips, 1920×1080, 30 fps), human-annotated with four annotation types on the same videos:
+As part of our GA announcement of our egocentric annotation API, we are releasing and open-sourcing 2 hours of human-annotated egocentric video.
+
+The annotations span four different categories:
 
 - Hand bounding box locations
 - Subtask annotations
-- Left-right hand actions, from [Ego4D's closed-vocab schema](https://ego4d-data.org/docs/benchmarks/hands-and-objects/)
-- Left-right hand captions
+- Left-right hand actions (from Ego4D's closed-vocab schema)
+- Left-right hand captions.
+
+We release the raw annotations as well as a code base for evaluation metrics which we believe better capture the inherent subjectivity of open vocabulary video description.
+
+### A new standard for Egocentric annotations
+
+Rich egocentric annotation is of increasing interest in the physical AI and robotics world, largely because of recent work demonstrating that it is highly effective as a scalable pretraining objective for world action models and vision language action models for [robotic manipulation tasks](https://www.dyna.co/dyna-2).
+
+Despite this increasing attention on egocentric data, it is still not perfectly well-understood which set of annotations serve as the best pretraining objectives or how best to measure the quality of synthetic labeling harnesses for these tasks.
+
+The most widely adopted egocentric benchmark, [WGO-Bench](https://macrodata.co/blog/annotating-robot-video-subtasks), has been a remarkably useful contribution to the ecosystem and has since become a de facto standard for evaluating the quality of a model or harness on describing the events in an egocentric video. Despite this it still has several gaps in coverage which we seek to address with this release.
+
+**Perceptron's EgoHome** benchmark addresses a few gaps left behind by the current state of egocentric benchmarking:
 
-It also introduces **permissible subsegment Jaccard**, a metric built for the fact that there are many valid ways to break down the granularity of a task.
+- We introduce benchmarking for four types of annotation all on the same video sources: hand box location, subtask breakdown, hand action classification (from the Ego4D closed vocab), and hand action freeform captions.
+- We provide a large number of annotations at the sub-second scale: 3,371 segments under 1 s and 773 under 0.5 s.
+- Finally, we introduce a new flexible metric we call permissible subsegment Jaccard. An LLM-as-judge based technique which accommodates the fact that there are many permissible ways to break down the granularity of a task.
 
 **Prompts and evaluation code:** [perceptron-ai-inc/egohome_public_release](https://github.com/perceptron-ai-inc/egohome_public_release)
 
@@ -55,34 +71,6 @@ It also introduces **permissible subsegment Jaccard**, a metric built for the fa
 
 ![Segment durations: EgoHome vs WGO-Bench](assets/fig_segment_duration_cdf.png)
 
-## Quick start
-
-```python
-from datasets import load_dataset
-
-ds = load_dataset("PerceptronAI/EgoHome", split="test")
-row = ds[0]
-row["video"]          # the clip (decoded with torchcodec)
-row["subtasks"]       # [{"start": 0.0, "end": 1.74, "label": "..."}, ...]
-row["hand_actions"]   # {"left": [...], "right": [...]}, Ego4D verb labels
-row["hand_captions"]  # {"left": [...], "right": [...]}
-row["hand_boxes"]     # [{"t": 0.0, "left": [x_min, y_min, x_max, y_max] | None, "right": ...}, ...]
-```
-
-To get the annotations without the videos, download `data/test/metadata.jsonl`.
-
-There is one row per clip. Times are seconds from the clip start. Boxes are `[x_min, y_min, x_max, y_max]`, normalized to 0–1000 (x over width, y over height).
-
-| Column | Contents |
-|---|---|
-| `video` / `file_name` | The mp4 |
-| `clip_id`, `source_video_id`, `clip_start_sec` | Clip identity (clips are cut from longer source videos) |
-| `duration_sec`, `fps`, `width`, `height` | Video properties |
-| `hand_boxes` | `[{t, left, right}]`, one entry per annotated frame; a hand is `null` when off screen |
-| `subtasks` | `[{start, end, label}]` |
-| `hand_actions` | `{left: [{start, end, label}], right: [...]}`; labels are Ego4D FHO verb classes |
-| `hand_captions` | `{left: [{start, end, label}], right: [...]}` |
-
 ## How different models stack up
 
 | System | Subtask | Hand action | Hand caption | Hand box | Failed | Timeline $ / video-hour | All tasks $ / video-hour |
@@ -154,6 +142,93 @@ Best in bold. ‡ Scored over all 137 clips; every other row covers the same 135
 
 </details>
 
+## EgoHome deep-dive
+
+### Annotation methodology
+
+We begin by synthetically labeling videos, and then having humans edit the initial annotations. We then had a member of our research team review every clip by hand. Due to the quality of our initial annotations and our proprietary review pipeline, we were able to get the overall average time to re-annotate a video down to 116 minutes per minute of video across all four annotation types.
+
+| Annotation type | Active minutes per minute of video | Agreement of the synthetic seed with the final human labels |
+|---|---:|---|
+| Hand boxes | 41.7 | 0.932 (centroid hit fraction) |
+| Subtasks | 17.1 | 0.766 (permissible subsegment Jaccard) |
+| Hand actions | 24.6 | 0.613 (permissible subsegment Jaccard) |
+| Hand captions | 32.6 | 0.688 (permissible subsegment Jaccard) |
+| All four | 115.9 | |
+
+#### Ensuring annotator quality
+
+Every annotator undergoes a calibration where they have 3 attempts to get high agreement with a clip labeled by a member of our research staff (the minimum performance criteria was determine by having multiple members of our research staff attempt the same clip and then choosing fair guidelines by which to include or exclude participants). After each attempt they are allowed to see what parts of their answer were marked as incorrect (but not able to see the ground truth), they are then allowed to re-attempt the calibration clip after seeing feedback.
+
+After their initial calibration, crowdsourced annotators then go through a 10-clip calibration against each other, where they are graded by how high their interannotator agreement is compared to their agreement with the original labels. Via this process we chose 5 annotators to serve as the annotators on our evaluation set, keeping lower performing members of our annotation pool to work on other tasks.
+
+### Four annotation types
+
+#### Subtask annotations
+
+Similar to WGO-Bench, we have open vocab descriptions of the high level goal of what task the subject is performing e.g. "scrubbing the stovetop". The average video has 14.0 segments. The distribution of durations is below:
+
+| Subtask segments | |
+|---|---:|
+| Segments | 1,924 |
+| Shortest | 0.16 s |
+| Median | 2.50 s |
+| Mean | 3.57 s |
+| 90th percentile | 7.34 s |
+| Longest | 55.1 s |
+
+| Duration | Segments |
+|---|---:|
+| < 0.5 s | 30 (1.6%) |
+| 0.5–1 s | 120 |
+| 1–2 s | 564 |
+| 2–5 s | 829 |
+| 5–10 s | 277 |
+| > 10 s | 104 |
+
+#### Hand box locations
+
+We annotate every hand box with human annotators at 10FPS then apply proprietary smoothing algorithms to extrapolate up to 30FPS. Here is the percentage of time the right and left hand are on screen in the corpus:
+
+| Hand on screen | % of frames |
+|---|---:|
+| Left | 96.7% |
+| Right | 96.6% |
+| Both | 94.3% |
+| Either | 99.0% |
+| Neither | 1.0% |
+| Boxes | 417,912 on 216,225 frames |
+
+#### Hand actions
+
+We have annotators choose an accurate hand action label from the closed vocabulary [Ego4D action schema](https://ego4d-data.org/docs/benchmarks/hands-and-objects/); there are 118 total actions to choose from (47 occur in the release).
+
+| Hand actions | Left | Right |
+|---|---:|---:|
+| Segments per clip | 17.3 | 19.4 |
+| Median duration | 1.45 s | 1.34 s |
+| Mean duration | 2.49 s | 2.39 s |
+| < 0.5 s | 171 | 248 |
+| < 1 s | 731 | 939 |
+| Distinct labels (both hands) | 47 | |
+| Time both hands have the exact same action | 42.7% of video time | 55.8% of time both hands are labeled |
+
+Most frequent verbs: hold (802), take (615), put (595), fold (427), move (310), arrange (257), adjust (227), smooth (214), lift (207), wash (163).
+
+#### Hand captions
+
+We have annotators write free-form hand captions describing what the hands are doing during the video.
+
+| Hand captions | Left | Right |
+|---|---:|---:|
+| Segments per clip | 18.4 | 20.1 |
+| Median duration | 1.55 s | 1.49 s |
+| Mean duration | 2.50 s | 2.38 s |
+| < 0.5 s | 145 | 179 |
+| < 1 s | 707 | 844 |
+| Distinct labels (both hands) | 3,948 | |
+| Time both hands have the exact same caption | 29.3% of video time | 35.3% of time both hands are labeled |
+
 ## Evaluation metrics
 
 ### Hand box metric: centroid hit fraction
@@ -227,90 +302,37 @@ Generic labels earn nothing on subtasks and captions. A constant "hold" earns 0.
 
 Label judge: gemini-3.5-flash, temperature 0, thinking off; prompt in the repository. Re-judging the same outputs moves a system's score by about ±0.01.
 
-## The annotations
-
-### Annotation methodology
-
-We begin by synthetically labeling videos, and then having humans edit the initial annotations. We then had a member of our research team review every clip by hand. Due to the quality of our initial annotations and our proprietary review pipeline, we were able to get the overall average time to re-annotate a video down to 116 minutes per minute of video across all four annotation types.
-
-| Annotation type | Active minutes per minute of video | Agreement of the synthetic seed with the final human labels |
-|---|---:|---|
-| Hand boxes | 41.7 | 0.932 (centroid hit fraction) |
-| Subtasks | 17.1 | 0.766 (permissible subsegment Jaccard) |
-| Hand actions | 24.6 | 0.613 (permissible subsegment Jaccard) |
-| Hand captions | 32.6 | 0.688 (permissible subsegment Jaccard) |
-| All four | 115.9 | |
-
-#### Ensuring annotator quality
-
-Every annotator undergoes a calibration where they have 3 attempts to get high agreement with a clip labeled by a member of our research staff (the minimum performance criteria was determine by having multiple members of our research staff attempt the same clip and then choosing fair guidelines by which to include or exclude participants). After each attempt they are allowed to see what parts of their answer were marked as incorrect (but not able to see the ground truth), they are then allowed to re-attempt the calibration clip after seeing feedback.
-
-After their initial calibration, crowdsourced annotators then go through a 10-clip calibration against each other, where they are graded by how high their interannotator agreement is compared to their agreement with the original labels. Via this process we chose 5 annotators to serve as the annotators on our evaluation set, keeping lower performing members of our annotation pool to work on other tasks.
-
-### Subtasks
-
-Similar to WGO-Bench, we have open vocab descriptions of the high level goal of what task the subject is performing e.g. "scrubbing the stovetop". The average video has 14.0 segments. The distribution of durations is below:
-
-| Subtask segments | |
-|---|---:|
-| Segments | 1,924 |
-| Shortest | 0.16 s |
-| Median | 2.50 s |
-| Mean | 3.57 s |
-| 90th percentile | 7.34 s |
-| Longest | 55.1 s |
-
-| Duration | Segments |
-|---|---:|
-| < 0.5 s | 30 (1.6%) |
-| 0.5–1 s | 120 |
-| 1–2 s | 564 |
-| 2–5 s | 829 |
-| 5–10 s | 277 |
-| > 10 s | 104 |
-
-### Hand boxes
-
-We annotate every hand box with human annotators at 10FPS then apply proprietary smoothing algorithms to extrapolate up to 30FPS. Here is the percentage of time the right and left hand are on screen in the corpus:
+Our metrics and prompts for scoring can be found here: https://github.com/perceptron-ai-inc/egohome_public_release
 
-| Hand on screen | % of frames |
-|---|---:|
-| Left | 96.7% |
-| Right | 96.6% |
-| Both | 94.3% |
-| Either | 99.0% |
-| Neither | 1.0% |
-| Boxes | 417,912 on 216,225 frames |
-
-### Hand actions
+And the Hugging Face dataset can be found here: https://huggingface.co/datasets/PerceptronAI/EgoHome
 
-We have annotators choose an accurate hand action label from the closed vocabulary [Ego4D action schema](https://ego4d-data.org/docs/benchmarks/hands-and-objects/); there are 118 total actions to choose from (47 occur in the release).
+## Using the dataset
 
-| Hand actions | Left | Right |
-|---|---:|---:|
-| Segments per clip | 17.3 | 19.4 |
-| Median duration | 1.45 s | 1.34 s |
-| Mean duration | 2.49 s | 2.39 s |
-| < 0.5 s | 171 | 248 |
-| < 1 s | 731 | 939 |
-| Distinct labels (both hands) | 47 | |
-| Time both hands have the exact same action | 42.7% of video time | 55.8% of time both hands are labeled |
+```python
+from datasets import load_dataset
 
-Most frequent verbs: hold (802), take (615), put (595), fold (427), move (310), arrange (257), adjust (227), smooth (214), lift (207), wash (163).
+ds = load_dataset("PerceptronAI/EgoHome", split="test")
+row = ds[0]
+row["video"]          # the clip (decoded with torchcodec)
+row["subtasks"]       # [{"start": 0.0, "end": 1.74, "label": "..."}, ...]
+row["hand_actions"]   # {"left": [...], "right": [...]}, Ego4D verb labels
+row["hand_captions"]  # {"left": [...], "right": [...]}
+row["hand_boxes"]     # [{"t": 0.0, "left": [x_min, y_min, x_max, y_max] | None, "right": ...}, ...]
+```
 
-### Hand captions
+To get the annotations without the videos, download `data/test/metadata.jsonl`.
 
-We have annotators write free-form hand captions describing what the hands are doing during the video.
+There is one row per clip. Times are seconds from the clip start. Boxes are `[x_min, y_min, x_max, y_max]`, normalized to 0–1000 (x over width, y over height).
 
-| Hand captions | Left | Right |
-|---|---:|---:|
-| Segments per clip | 18.4 | 20.1 |
-| Median duration | 1.55 s | 1.49 s |
-| Mean duration | 2.50 s | 2.38 s |
-| < 0.5 s | 145 | 179 |
-| < 1 s | 707 | 844 |
-| Distinct labels (both hands) | 3,948 | |
-| Time both hands have the exact same caption | 29.3% of video time | 35.3% of time both hands are labeled |
+| Column | Contents |
+|---|---|
+| `video` / `file_name` | The mp4 |
+| `clip_id`, `source_video_id`, `clip_start_sec` | Clip identity (clips are cut from longer source videos) |
+| `duration_sec`, `fps`, `width`, `height` | Video properties |
+| `hand_boxes` | `[{t, left, right}]`, one entry per annotated frame; a hand is `null` when off screen |
+| `subtasks` | `[{start, end, label}]` |
+| `hand_actions` | `{left: [{start, end, label}], right: [...]}`; labels are Ego4D FHO verb classes |
+| `hand_captions` | `{left: [{start, end, label}], right: [...]}` |
 
 ## Citation