mantelmoritz/CorridorKey-Synthetic-Streamers

CorridorKey Synthetic Streamers

275 AI-generated livestream / facecam video clips in front of green screens, with per-frame alpha mattes and foreground colours computed by CorridorKey 2.0 at native resolution.

Why this dataset exists

Real green-screen footage of streamers is hard to come by: public stock footage shows actors in studios, while a real facecam setup looks very different. Typical facecam conditions:

• cheap webcams or mirrorless cameras with noise and 4:2:0 chroma
• pull-up screens that are too small, wrinkled or unevenly lit
• RGB lighting that spills onto hair and headphones
• gaming headsets, boom-arm microphones, glasses, mugs, energy drinks and desk clutter
• people leaning in and out of frame

This dataset was generated to train and evaluate a real-time green-screen keyer for OBS: the CorridorKey Real-Time Edge Refiner, a tiny network that restores native-resolution hair detail on top of CorridorKey 2.0 when CorridorKey runs at reduced resolution for live use.

It is useful for:

• Training and evaluating video matting / keying models on facecam-style content (hair, headsets, microphones, glasses, transparent cups, motion).
• Distillation: the labels are the output of a large, slow offline keyer (CorridorKey 2.0 at full resolution), so they are a natural target for small real-time students.
• Robustness tests: dark, noisy, spill-heavy, partially covered and badly lit screens on purpose.
• Re-compositing: straight foreground plus alpha lets you place the subjects on any background or on new, procedurally generated green screens (which gives exact ground truth for the re-composited frames).

What is in it

Clips	275 (233 train / 42 test), about 5 s each, 24 fps (one 30 fps clip), 1411 s in total
Resolution	2560×1440 (134 clips), 1920×1080 (140), 1280×720 (1)
Labelled frames	40 consecutive frames per clip (frames 12–51), 11,000 frames in total
Labels	alpha (16-bit PNG) and straight foreground colour (sRGB, lossless WebP) at the clip's native resolution
Backgrounds	251 clips with a green screen, 24 clips filmed in a real room without a screen (see notes)

Generator models

All people, rooms and objects are synthetic. No real person was filmed. Any resemblance to real people is coincidental. The clips were generated from text prompts (included per clip) with these video models:

Model	Clips
Kling 4.0 (Kuaishou)	134
MiniMax H3 (MiniMax)	134
Seedance 2.5 (ByteDance)	4
Kling 3.0 Turbo (Kuaishou)	1
Google Veo 3.1 Lite (Google)	1
Wan 3.0 (Alibaba)	1

Series

Series	Clips	Content
pilot	7	first comparison of generator models on the same prompt
s1	32	curly / long / frizzy hair, headphones, microphones, many screen colours and lighting setups
s2	120	wider variety of people, hair styles, accessories, desks, lighting; 24 real-room clips
s3	116	harder cases: foreground objects (cups, cans, hands, props), more screen types, partially covered screens, people standing up and leaving the frame

Layout

metadata.jsonl                     one row per clip (see below)
clips/<id>.mp4                     the generated video, unchanged
labels/<id>/alpha/<frame>.png      16-bit grayscale alpha, <frame> = absolute frame index in the clip (000012 ...)
labels/<id>/fg/<frame>.webp        straight (un-premultiplied) foreground, sRGB 8-bit, lossless; 0 where alpha = 0
images/                            figures of this card

metadata.jsonl fields:

Field	Meaning
file_name	path of the clip
id	clip id (<series>_<number>[_room]_<generator>)
series	pilot, s1, s2, s3
generator	video model that generated the clip
split	train or test (the test clips were never used for training the edge refiner)
background	green screen or real room (no green screen)
width, height, fps, num_frames	video properties
label_first_frame, label_num_frames	labelled frame range (12, 40)
label_notes	known label problems of this clip (empty if none are known)
prompt	the full generation prompt

Loading a labelled frame

import json, cv2, numpy as np

row = json.loads(open("metadata.jsonl").readline())
cid, f = row["id"], row["label_first_frame"] + 5

# The labels were computed on frames decoded with ffmpeg (-pix_fmt rgb24). OpenCV's decoder uses a slightly
# different YUV->RGB conversion (about 1-2 / 255 mean difference); for exact pixels use ffmpeg:
#   ffmpeg -i clips/<id>.mp4 -vf "select=gte(n\,12)" -fps_mode passthrough -frames:v 40 -pix_fmt rgb24 -start_number 12 %06d.png
cap = cv2.VideoCapture(row["file_name"])
cap.set(cv2.CAP_PROP_POS_FRAMES, f)
frame = cap.read()[1][:, :, ::-1] / 255.0                                         # RGB, sRGB-encoded

alpha = cv2.imread(f"labels/{cid}/alpha/{f:06d}.png", cv2.IMREAD_UNCHANGED) / 65535.0
fg = cv2.imread(f"labels/{cid}/fg/{f:06d}.webp")[:, :, ::-1] / 255.0               # straight sRGB foreground

background = np.zeros_like(fg); background[..., 2] = 1.0                           # any new background
composite = fg * alpha[..., None] + background * (1.0 - alpha[..., None])           # (compositing in sRGB for brevity)

How the labels were made

Each clip was decoded to lossless frames and keyed with the official CorridorKey 2.0 offline pipeline (infer.py, native resolution, automatic windowing). Alpha and foreground are the direct outputs, only converted to 16-bit PNG (alpha) and 8-bit sRGB (foreground; the original output is scene-linear float).

Important limitations

• These are pseudo-labels, not ground truth. They are as good as CorridorKey 2.0 is on the clip. That is very good on clean screens, but it has visible errors. Typical ones:

  • holes in dark headphone ear cups
  • partially keyed chairs
  • soft, slightly blurry hair at native resolution
  • parts of desks and props kept or removed inconsistently

  Use them as a strong teacher signal, not as a benchmark of absolute matting accuracy.

• Green clothing: in 12 clips of series 3 the person wears green clothing in front of the green screen, and CorridorKey keys parts of the clothing away. These clips are marked in label_notes and were excluded from training the edge refiner.

• Real-room clips (background = real room): CorridorKey is a green-screen keyer. Without a screen it keeps almost the whole frame as foreground, so these labels are not background-removal labels. They are included as hard negatives / robustness material only and were excluded from training.

• Generated video artefacts: hair and fine detail are produced by video generators. They can show smearing, temporal shimmer and occasional anatomical errors. They are realistic enough for keying, but not real camera footage. The test split of a model trained here should always be complemented by real footage.

• Only 40 frames per clip are labelled. The full clip is included for context and temporal models.

License and attribution

• Dataset (clips, labels, metadata): CC BY-NC-SA 4.0. Non-commercial use only, share alike, attribution required.
• Labels made with CorridorKey 2.0 by Corridor Digital LLC (model designed and trained by Zachary Harris in collaboration with Niko Pueringer and Corridor Digital): https://huggingface.co/nikopueringer/CorridorKey2.0, https://corridorkey.com. CorridorKey 2.0 itself is licensed under the Corridor Key Licence (a copy is included as LICENSE_CorridorKey.txt). This dataset does not contain CorridorKey code or weights.
• This is an independent community project. It is not affiliated with, sponsored by or endorsed by Corridor Digital LLC.

Citation

@misc{corridorkey_synthetic_streamers_2026,
  title  = {CorridorKey Synthetic Streamers: AI-generated facecam green-screen clips with CorridorKey 2.0 mattes},
  author = {Mantel, Moritz},
  year   = {2026},
  howpublished = {\url{https://huggingface.co/datasets/mantelmoritz/CorridorKey-Synthetic-Streamers}}
}