pictograph/pg-handpose

View on Pictograph · Pictograph Research · Keypoint Detection · RF-DETR Keypoint Preview · FP32 · Creative Commons Attribution-NonCommercial-ShareAlike 4.0

Overview

This is a keypoint detection model built on RF-DETR Keypoint Preview, trained on the Hand Keypoints dataset. It expects 576x576 input and runs in FP32 precision. It finds each hand and its 21 keypoints: the Skeleton section lists them in output order, with the joints each connects to. The ONNX weights are in this repo; run them with the Pictograph SDK (below) or open the model on Pictograph to test it in-browser, call the hosted API, or deploy it as an always-on endpoint.

Performance

Headline: 81.8% OKS.

Metric	Value
OKS	81.8%
OKS@50	92.7%
OKS@75	84.0%
Recall	85.1%

OKS is COCO's keypoint AP: the average precision over object keypoint similarity (OKS) thresholds 0.50 to 0.95. OKS@50 and OKS@75 are the AP at one threshold; Recall is the average recall over 0.50 to 0.95.

Skeleton

Each object the model finds is one instance: its keypoints share an instance_id, are named for their joint and carry their own confidence. The keypoint index is the model's output order.

#	Keypoint	Connects to
0	wrist	thumb_cmc, index_finger_mcp, pinky_mcp
1	thumb_cmc	wrist, thumb_mcp
2	thumb_mcp	thumb_cmc, thumb_ip
3	thumb_ip	thumb_mcp, thumb_tip
4	thumb_tip	thumb_ip
5	index_finger_mcp	wrist, index_finger_pip, middle_finger_mcp
6	index_finger_pip	index_finger_mcp, index_finger_dip
7	index_finger_dip	index_finger_pip, index_finger_tip
8	index_finger_tip	index_finger_dip
9	middle_finger_mcp	index_finger_mcp, middle_finger_pip, ring_finger_mcp
10	middle_finger_pip	middle_finger_mcp, middle_finger_dip
11	middle_finger_dip	middle_finger_pip, middle_finger_tip
12	middle_finger_tip	middle_finger_dip
13	ring_finger_mcp	middle_finger_mcp, ring_finger_pip, pinky_mcp
14	ring_finger_pip	ring_finger_mcp, ring_finger_dip
15	ring_finger_dip	ring_finger_pip, ring_finger_tip
16	ring_finger_tip	ring_finger_dip
17	pinky_mcp	ring_finger_mcp, wrist, pinky_pip
18	pinky_pip	pinky_mcp, pinky_dip
19	pinky_dip	pinky_pip, pinky_tip
20	pinky_tip	pinky_dip

Training

Setting	Value
Base dataset	Hand Keypoints
Dataset size	26,677 images
Input size	576x576
Epochs	12
Batch size	4
Learning rate	0.0001
Model size	preview
Precision	FP32
Version	1.0.0

Use this model

The weights ship as ONNX (pg-handpose.onnx) with a config.json. The Pictograph SDK applies the model's exact pre/post-processing for you:

Pictograph SDK

pip install "pictograph[inference]"

from huggingface_hub import hf_hub_download
from pictograph import load_model, KeypointModel

onnx = hf_hub_download("pictograph/pg-handpose", "pg-handpose.onnx")
config = hf_hub_download("pictograph/pg-handpose", "config.json")

model: KeypointModel = load_model(onnx, config, task="keypoint_detection")
result = model.predict("photo.jpg")   # KeypointResult
for obj in result.instances:           # one list of joints per object
    print(f"instance {obj[0].instance_id}: {len(obj)} joints")
    for j in obj:                      # j.name IS the joint class
        print(f"  {j.name} ({j.keypoint.x:.0f},{j.keypoint.y:.0f}) {j.confidence:.2f}")

Example output (illustrative; your exact values depend on the image):

instance 1: 21 joints
  wrist (512,143) 0.97
  thumb_cmc (498,131) 0.96
instance 2: 21 joints
  wrist (188,160) 0.93

Raw ONNX Runtime (no SDK)

from huggingface_hub import hf_hub_download
import onnxruntime as ort

onnx = hf_hub_download("pictograph/pg-handpose", "pg-handpose.onnx")
sess = ort.InferenceSession(onnx, providers=["CPUExecutionProvider"])
inp = sess.get_inputs()[0]
print(inp.name, inp.shape)   # NCHW float32; input size + normalization in config.json
# outputs = sess.run(None, {inp.name: preprocessed_batch})
# Decoding is architecture-specific; the Pictograph SDK (above) does it for you.

• Test it in-browser, call the hosted API, or deploy an endpoint: open the model on Pictograph.
• SDK + local inference guide: docs.

License

Released under Creative Commons Attribution-NonCommercial-ShareAlike 4.0.

Citation

This model uses the RF-DETR Keypoint Preview architecture:

@inproceedings{robinson2026rfdetr,
  title     = {RF-DETR: Real-Time Detection Transformer},
  author    = {Robinson, Isaac and Robicheaux, Peter and Popov, Matvei and Ramanan, Deva and Peri, Neehar},
  booktitle = {International Conference on Learning Representations (ICLR)},
  year      = {2026},
  url       = {https://arxiv.org/abs/2511.09554}
}

---

Trained and published with Pictograph - annotate, train, and deploy from one API.