SeetaPsych Face Hub
Community open-source face modules for SeetaPsych
Usage
This project is already included in the seetapsych-lib default configuration. Download and use it via seetapsych-manager download.
For usage, refer to SeetaPsych.
You can additionally add this algorithm module using the following methods.
WebUI
Run seetapsych-webui with the --dirs argument to use it.
seetapsych-webui --files seetapsych_face_hub/modules/insightface/retinaface.yml
Programmatic Usage
Add the following code in your program to use this algorithm module.
from seetapsych_lib.runtime.factory import Factory
from seetapsych_lib.runtime.pipeline import Pipeline
factory = Factory()
factory.load_file_modules("seetapsych_face_hub/modules/insightface/retinaface.yml")
pipeline = Pipeline(factory, ...)
pipeline.add_attributes("face/detection")
Module Catalog
| Module YML | Packages |
|---|---|
| insightface/retinaface.yml | FaceDetection-RetinaFace(InsightFace) |
| insightface/arcface.yml | FaceFeature-ArcFace(InsightFace) |
| mediapipe.yml | FaceDetection-MediaPipe, FaceMesh-MediaPipe |
| retinaface.yml | FaceDetection-RetinaFace(PyTorch) |
InsightFace (RetinaFace)
InsightFace Buffalo RetinaFace detector: face bounding box + 5-point landmark detection with 500M / 2.5G / 10G FLOPs model variants.
Module config: insightface/retinaface.yml
| Package Name | Provides Attributes | Requires Attributes |
|---|---|---|
| FaceDetection-RetinaFace(InsightFace) | face/detection, face/landmarks |
(none) |
Description: InsightFace RetinaFace detecting face boxes + 5-point landmarks with configurable input size; 3 accuracy/speed model tiers
Parameters
| Name | Type | Default | Description & Tuning |
|---|---|---|---|
| input_size | number[] | [640, 640] | Network input [width, height] in pixels. Larger sizes improve small-face recall but slow inference. Square inputs are recommended. |
Models
| Name | Recommended | Notes |
|---|---|---|
| det_500m.onnx | ✓ | |
| det_2.5g.onnx | ||
| det_10g.onnx |
Output Attributes
InsightFace (ArcFace)
InsightFace Buffalo face feature extraction via ArcFace models producing L2-normalized 512-dim embeddings.
Module config: insightface/arcface.yml
| Package Name | Provides Attributes | Requires Attributes |
|---|---|---|
| FaceFeature-ArcFace(InsightFace) | face/feature |
face/landmarks |
Description: Extract L2-normalized 512-dim face embeddings for recognition, clustering, or similarity search from 5-point aligned faces
Parameters: (none)
Models
| Name | Recommended | Notes |
|---|---|---|
| w600k_r50.onnx | ✓ | |
| w600k_mbf.onnx |
Output Attributes
face/feature— spec.
MediaPipe
MediaPipe Tasks face detection and 468-point 3D face mesh landmarking via BlazeFace short/full-range models.
Module config: mediapipe.yml
| Package Name | Provides Attributes | Requires Attributes |
|---|---|---|
| FaceDetection-MediaPipe | face/detection |
(none) |
| FaceMesh-MediaPipe | face/mesh |
face/detection |
Package: FaceDetection-MediaPipe
Description: Face bounding box detection via MediaPipe BlazeFace detector with short/full-range/sparse model variants
Parameters
| Name | Type | Default | Description & Tuning |
|---|---|---|---|
| running_mode | selection (IMAGE, VIDEO, LIVE_STREAM) | IMAGE | Task execution mode. IMAGE for single static frames; VIDEO/LIVE_STREAM for time-series smoothing across frames. |
| min_detection_confidence | number | 0.5 | Minimum confidence score for a detection box to be returned. Lower values catch more faces at the cost of false positives. |
| min_suppression_threshold | number | 0.3 | IoU threshold for non-maximum suppression. Higher values keep more overlapping boxes; lower values prune aggressively. |
Models
| Name | Recommended | Notes |
|---|---|---|
| blaze_face_full_range.tflite | ✓ | |
| blaze_face_full_range_sparse.tflite | ||
| blaze_face_short_range.tflite |
Output Attributes
face/detection— spec.
Package: FaceMesh-MediaPipe
Description: 468-point normalized 3D face mesh landmark extraction from detected face boxes with optional blendshapes and pose matrices
Parameters
| Name | Type | Default | Description & Tuning |
|---|---|---|---|
| running_mode | selection (IMAGE, VIDEO, LIVE_STREAM) | IMAGE | Task execution mode. VIDEO/LIVE_STREAM enable temporal landmark smoothing when num_faces = 1. |
| num_faces | integer | 1 | Maximum number of faces to landmark simultaneously. Smoothing is only active when num_faces is 1. |
| min_face_detection_confidence | number | 0.5 | Minimum confidence for the built-in face detector to succeed when initializing landmark placement. |
| min_face_presence_confidence | number | 0.5 | Minimum confidence that a face is present in the frame before landmark output; filters intermittent dropouts in video. |
| min_tracking_confidence | number | 0.5 | Minimum tracking confidence to reuse the last known face position vs. re-running full detection. Higher values reduce drift but trigger re-detection more often. |
| output_face_blendshapes | boolean | false | Whether to output 52 face blendshape coefficients for 3D avatar/AR rendering. Adds minor compute overhead. |
| output_facial_transformation_matrixes | boolean | false | Whether to output 4x4 facial pose transformation matrices for rigid head pose estimation. |
Models
| Name | Recommended | Notes |
|---|---|---|
| face_landmarker.task | ✓ |
Output Attributes
face/mesh— spec.
References
Metadata
Release files for seetapsych-face-hub 0.0.3.post1
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| seetapsych_face_hub-0.0.3.post1.tar.gz | 283.4 kB | With provenance | Details for seetapsych_face_hub-0.0.3.post1.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| seetapsych_face_hub-0.0.3.post1-py3-none-any.whl | Python 3 | none | any | Details for seetapsych_face_hub-0.0.3.post1-py3-none-any.whl |
Total release size: 346.8 kB
Release files / seetapsych_face_hub-0.0.3.post1.tar.gz
| Download URL | seetapsych_face_hub-0.0.3.post1.tar.gz |
|---|---|
| Size | 283.4 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
e6e47ee5540e482a158e077120d0c7e42d39ee794c26d12c278fb6313eebf43b
|
|
BLAKE2b-256 checksum How to use checksums |
afd6cef96f545b8ab5d98b38d8d8b3542d3da517591676a4acabbc9c4fcd1dc0
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.14
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Sep 3, 2026.
Transparency logRelease files / seetapsych_face_hub-0.0.3.post1-py3-none-any.whl
| Download URL | seetapsych_face_hub-0.0.3.post1-py3-none-any.whl |
|---|---|
| Size | 63.4 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
ea97fc63f9fe5c7a56ab79b61549b426ed86986659bae2d9f0b4f74d7115a625
|
|
BLAKE2b-256 checksum How to use checksums |
ec81e4216eab62c250a4e6478163eb15cef0e28a698b2c75ad5b03830d9ba434
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.14
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Sep 3, 2026.
Transparency logSecurity for seetapsych-face-hub 0.0.3.post1
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
Full provenance
Both files in this release have been uploaded with provenance metadata.
Attestation for seetapsych_face_hub-0.0.3.post1.tar.gz
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Sep 3, 2026.
Transparency logA note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.