louisye/social-robotics-attention

Social Robotics: Attention / Engagement (03a)

Does the bystander look at the camera-wearer? Per-bystander gaze + head-pose engagement around each task.

One layer of the Social-Affective Filter (SAF) — dehydrated social-signal metadata extracted from egocentric (first-person) video so robots can learn to read human reactions. No raw pixels and no audio. Each row is one source video, keyed by video_id; rehydrate against your own legally-obtained Ego4D copies (below).

• Rows: 23 — videos in the 50-clip evaluation slice for which this layer produced a measured signal (videos it could not measure are excluded from this per-layer dataset; the layers still join 1:1 on video_id).
• Signal: Per-bystander visual attention toward the POV actor (gaze raycast to the camera / the wearer's hands), scored 0–1 over an adaptive temporal trace.
• Method: L2CS-Net 3D gaze + MediaPipe FaceLandmarker head pose, sampled adaptively at 8→16 FPS.

⚠️ Read this first — interpretation caveats

• Trace timestamps (in the *_raw column) are not uniformly spaced (adaptive stride); resample onto a fixed-dt grid before frequency-domain analysis.
• A missing row would mean no bystander face was trackable; such rows are excluded from this per-layer dataset.
• Egocentric footage is legitimately low-yield (small/sparse bystander faces, heavy camera motion); we publish honest measurements only, never fabricated zeros.

Columns

Identity & manifest (shared across all SAF datasets)

column	type	meaning
video_id	string	Ego4D source-clip UUID. The rehydration key — map back to your own legally-obtained Ego4D copy (<video_id>.mp4).
source_dataset	string	Origin corpus (ego4d).
task_labels	string	Comma-joined VLM task label(s) — the activity the camera-wearer performed.
duration_sec	float	Source clip duration (seconds).
fps	float	Source clip frame rate.

Attention / Engagement signal

column	type	meaning
attention_num_bystanders_tracked	float (int)	Number of bystanders for whom a gaze trace was produced.
attention_mean_attention_all_persons	float (0–1)	Mean average_attention_score across tracked bystanders. Higher = more visually focused on the wearer/task.
attention_any_person_engaged	bool	True if any bystander crossed the engagement threshold.
attention_per_person_raw	JSON string	Per-bystander detail: person_id, average_attention_score, peak_engagement_timestamp_sec, attention_variance, sustained_engagement_sec, is_engaged, gaze_target_classification (Camera/POV_Actor_Hands/Unknown), and attention_trace of {t, score, pitch_rad, yaw_rad, target} (head-pose Euler angles in radians).
attention_processing_meta_model_used	string	Gaze model id (e.g. l2cs_net_3d_gaze).
attention_processing_meta_sampling_fps_effective	float	Baseline sampling rate (8 FPS).
attention_processing_meta_sampling_fps_burst	float	Boosted rate during fast attention transitions (16 FPS).
attention_processing_meta_sampling_strategy	string	e.g. adaptive_8_to_16_fps.

The *_raw JSON column

The *_raw column holds the full nested per-task / per-person detail as a JSON string. Parse it with:

import json, pandas as pd
df = pd.read_parquet("hf://datasets/louisye/social-robotics-attention/social_metadata.parquet")
raw_col = next(c for c in df.columns if c.endswith("_raw"))
detail = json.loads(df.iloc[0][raw_col])

How to load

import pandas as pd
df = pd.read_parquet("hf://datasets/louisye/social-robotics-attention/social_metadata.parquet")
# or:  from datasets import load_dataset;  ds = load_dataset("louisye/social-robotics-attention")

Rehydration — mapping back to video

video_id is the Ego4D clip UUID. With your own licensed Ego4D copy, the file is <video_id>.mp4; timestamps in the *_raw columns index into that clip. A helper (rehydrate_dataset.py) is included. We never redistribute source media — obtain Ego4D under its own license.

Provenance

Generated by the SAF pipeline (export_metadata.json records schema_version + pipeline_git_sha). Headers are the descriptive layer name + metric; the pipeline's internal 03a_ layer-id prefix is stripped at publish time. License MIT (this metadata only; Ego4D videos remain under the Ego4D license).