Dark Mode

Skip to content

Navigation Menu

Sign in
Appearance settings

Search code, repositories, users, issues, pull requests...

Provide feedback

We read every piece of feedback, and take your input very seriously.

Saved searches

Use saved searches to filter your results more quickly

Sign up
Appearance settings

blakeliu/CVHumanImgKit

Repository files navigation

FaceImageKit

PythonShi Xian De Ren Lian Tu Xiang Xiang Guan Suan Fa Pipeline

Install

python

example:

uv venv --python 3.12

deploy

cpu

uv pip install -e .[cpu]

cuda

uv pip install -e .[cuda]

develop

cpu

uv sync --extra cpu

cuda

uv sync --extra cuda

dev

uv sync --extra dev --extra cpu

Models

Face Dtection

scrfd

Models accuracy on WiderFace benchmark:

Model Easy Medium Hard
scrfd_10g_gnkps 95.51 94.12 82.14
scrfd_2.5g_gnkps 93.57 91.70 76.08
scrfd_500m_gnkps 88.70 86.11 63.57

Lai Yuan https://github.com/SthPhoenix/InsightFace-REST/

runtime

  • onnxruntime(cpu, cuda)
  • ncnn
  • opencv
  • TensorRT

Face Landmark

datasets: Lapa134(Lapa106 + 28)

PFLD

Xiang Mu Lai Yuan :https://gold-one.coding.net/p/BeautyFace/d/TFace_Landmark/git Mo Xing Shu Chu Da Xiao : $134\times 2$ runtime

  • onnxruntime(cpu)
  • ncnn
  • TensorRT

rtmface

Xiang Mu Lai Yuan https://blakeliu.coding.net/p/face/d/mmpose/git/tree/master/configs/face_2d_keypoint/rtmpose/lapa Mo Xing Shu Chu Da Xiao : $134\times 2$

Model NME
rtmpose-m-ort-lapa134 0.0288
rtmpose-s-ort-lapa134 0.0258
  • onnxruntime(cpu, cuda)
  • opencv
  • ncnn
  • TensorRT

Face Segmentation

ppliteseg

Xiang Mu Lai Yuan :https://github.com/tfrbt/FaceSeg.git

  • label map Lei Bie Shu Liang 32
id class
0 'background'
1 'skin'
2 'cheek'
3 'chin'
4 'ear'
5 'helix'
6 'lobule'
7 'bottom_lid'
8 'pupil'
9 'iris'
10 'sclera'
11 'tear_duct'
12 'top_lid'
13 'eyebrow'
14 'forhead'
15 'frown'
16 'hair'
17 'temple'
18 'jaw'
19 'beard'
20 'inferior_lip'
21 'oral comisure'
22 'superior_lip'
23 'teeth'
24 'neck'
25 'nose'
26 'ala_nose'
27 'bridge'
28 'nose_tip'
29 'nostril'
30 'DU26'
31 'sideburns'
Model val IOU val Dice
face_seg_ppliteseg_t 0.727 0.834
  • onnxruntime(cpu)
  • ncnn
  • TensorRT
  • label map Lei Bie Shu Liang 12
id class
0 "background" # 0.Bei Jing
1 "skin" # 1.Pi Fu
2 "eye" # 2.Yan Jing
3 "pupil" # 3.Tong Kong
4 "bottom_lid", # 4.Xia Yan Pi
5 "top_lid", # 5.Shang Yan Pi
6 "eyebrow" # 6.Mei Mao
7 "hair" # 7.Tou Fa
8 "superior_lip" # 8Shang Zui Chun
9 "teeth" # 9.Ya Chi
10 "inferior_lip" # 10.Xia Zui Chun
11 "nose" # 11.Bi Zi
Model val IOU val Dice
face12_seg_pplitesegb 0.774 0.869
  • onnxruntime(cpu)
  • ncnn
  • TensorRT

About

Modelling functions such as face, body or hand corner detection, key points, and segmentation based on onnx, openvino, and TRT

Resources

Readme

Stars

Watchers

Forks

Releases

No releases published

Packages

Contributors

Languages