Datasets:

RunCam
/

runcam-feed-camera-egocentric-rgb-imu

video_id stringclasses 6 values	source_file_name stringclasses 6 values	camera stringclasses 1 value	firmware_version stringclasses 1 value	video_resolution stringclasses 1 value	video_fps int64 60 60	video_codec stringclasses 1 value	nominal_bitrate_mbps int64 18 18	measured_video_bitrate_mbps float64 18.6 18.7	video_duration_seconds float64 26.9 68.7	container_duration_seconds float64 26.9 68.7 ⌀	video_file_size_bytes int64 63.2M 162M	start_timestamp_local timestamp[s]date 2018-08-04 15:05:21 2018-08-04 15:12:54	imu_sensor stringclasses 1 value	imu_axes int64 6 6	imu_sampling_rate_hz int64 800 800	imu_timestamp_scale_seconds float64 0 0	imu_sample_count int64 21.3k 54.8k	imu_duration_seconds float64 26.6 68.5 ⌀	sync_offset_ms null	subject_id stringclasses 1 value	session_id stringclasses 1 value	task stringclasses 6 values	environment stringclasses 1 value	annotation stringclasses 6 values	imu_metadata_track_verified bool 1 class	notes stringclasses 1 value
V01	V01_0023_20180804150521.MP4	RunCam Feed Camera	FIRMWARE_0.1.0	1920x1080	60	H.265 / HEVC	18	18.749	68.683	68.736	162,386,449	2018-08-04T15:05:21	ICM-42607	6	800	0.00125	54,788	68.48375	null	S01	Session_001	Taking items out of a zippered pouch	Indoor / Office Desk	Seated first-person POV recording showing the operator taking items out of a zippered pouch one by one on a desk.	false	The GCSV file identifies the associated video as V01_0023_20180804150521.mp4. The 800 Hz sampling rate was calculated from the GCSV timestamp scale and consecutive sample timestamps.
V02	V02_0024_20180804150640.MP4	RunCam Feed Camera	FIRMWARE_0.1.0	1920x1080	60	H.265 / HEVC	18	18.74	65.217	null	154,106,964	2018-08-04T15:06:40	ICM-42607	6	800	0.00125	52,076	null	null	S01	Session_001	Putting items back into a zippered pouch	Indoor / Office Desk	Seated first-person POV recording showing the operator putting items back into a zippered pouch one by one on a desk.	null	null
V03	V03_0025_20180804150758.MP4	RunCam Feed Camera	FIRMWARE_0.1.0	1920x1080	60	H.265 / HEVC	18	18.714	55.25	null	130,374,536	2018-08-04T15:07:58	ICM-42607	6	800	0.00125	44,047	null	null	S01	Session_001	Finding a purple marker and writing on paper	Indoor / Office Desk	Seated first-person POV recording showing the operator taking out a purple marker from a zippered pouch and writing on an A4 sheet of paper.	null	null
V04	V04_0026_20180804150916.MP4	RunCam Feed Camera	FIRMWARE_0.1.0	1920x1080	60	H.265 / HEVC	18	18.729	51.033	null	120,523,172	2018-08-04T15:09:16	ICM-42607	6	800	0.00125	40,713	null	null	S01	Session_001	Finding scissors and cutting handwritten content	Indoor / Office Desk	Seated first-person POV recording showing the operator searching for scissors from a zippered pouch, then cutting out the handwritten content on an A4 sheet of paper.	null	null
V05	V05_0027_20180804151041.MP4	RunCam Feed Camera	FIRMWARE_0.1.0	1920x1080	60	H.265 / HEVC	18	18.637	26.867	26.88	63,165,407	2018-08-04T15:10:41	ICM-42607	6	800	0.00125	21,307	26.633	null	S01	Session_001	Folding pants on a table	Indoor / Office Desk	Standing first-person POV recording showing the operator folding a pair of pants on a desk.	null	null
V06	V06_0029_20180804151254.MP4	RunCam Feed Camera	FIRMWARE_0.1.0	1920x1080	60	H.265 / HEVC	18	18.681	33.5	33.536	78,903,278	2018-08-04T15:12:54	ICM-42607	6	800	0.00125	26,602	33.251	null	S01	Session_001	Packing items into a bag on a table	Indoor / Office Desk	Standing first-person POV recording showing the operator packing items into a bag on a desk.	null	null

RunCam Feed Camera - Egocentric RGB + IMU Sample Dataset

A small sample dataset captured with the RunCam Feed Camera for egocentric video and synchronized motion-sensor workflows.

Capture Device

Video: H.265 MP4, 1920x1080, 60 fps for V01-V06
Nominal video bitrate: 18 Mbps
Horizontal field of view: 126 degrees
Device weight: approximately 26 g
IMU: ICM-42607, 6-axis
IMU sampling rate: 800 Hz for the recordings in this sample
Firmware reported in the GCSV files: FIRMWARE_0.1.0

Data Included

Each sample contains one RGB video (RGB.mp4), one motion file (IMU.gcsv), and one recording metadata file (metadata.json). The repository also includes catalog.csv and examples/read_imu.py.

Repository Structure

catalog.csv
examples/
  read_imu.py
samples/
  V01/ ... V06/
    RGB.mp4
    IMU.gcsv
    metadata.json

IMU Data Format

The GCSV files contain a metadata header followed by comma-separated samples:

t,gx,gy,gz,ax,ay,az

t: elapsed time in seconds
gx, gy, gz: gyroscope readings
ax, ay, az: accelerometer readings
tscale=0.00125 seconds per sample, corresponding to 800 Hz
IMU identifier: icm42607

Timestamp and Synchronization

Video timestamps are represented by the MP4 container. IMU timestamps are elapsed-time values in the GCSV file. The paired files were captured during the same device session, and GCSV metadata identifies the corresponding video filename. The nominal IMU interval is 1.25 ms. Exact cross-modal sync offset has not been independently calibrated.

Sample Description

ID	Activity	Environment	Viewpoint
V01	Taking items out of a zippered pouch	Indoor / office desk	Seated first-person POV
V02	Putting items back into a zippered pouch	Indoor / office desk	Seated first-person POV
V03	Finding a purple marker and writing on paper	Indoor / office desk	Seated first-person POV
V04	Finding scissors and cutting handwritten content	Indoor / office desk	Seated first-person POV
V05	Folding pants on a table	Indoor / office desk	Standing first-person POV
V06	Packing items into a bag on a table	Indoor / office desk	Standing first-person POV

Python Example

from pathlib import Path
import pandas as pd

path = Path("samples/V01/IMU.gcsv")
text = path.read_text(encoding="utf-8", errors="replace")
header_end = text.index("t,gx,gy,gz,ax,ay,az")
metadata = {}
for line in text[:header_end].splitlines():
    if "=" in line:
        key, value = line.split("=", 1)
        metadata[key.strip()] = value.strip()
imu = pd.read_csv(path, skiprows=len(text[:header_end].splitlines()))
print(metadata)
print(imu.head())
print("Samples:", len(imu))
print("Estimated rate:", round(1 / float(metadata["tscale"])), "Hz")

Intended Use

This sample is intended for research, prototyping, sensor-format inspection, egocentric video experiments, and evaluation of lightweight RGB + IMU capture workflows.

Limitations

This is a small indoor sample and is not a complete robotics sensing suite. It does not provide RGB-D, LiDAR, force, torque, or tactile channels. Video-IMU synchronization should be independently validated for safety-critical or high-precision use.

Contact

For technical questions or evaluation discussions, please contact RunCam.

Downloads last month: 164

Number of rows:

Total file size:

717 MB