Metadata-Version: 2.1
Name: MediapipeOpenvino
Version: 0.0.1
Summary: Towards a subset reimplmention of Mediapipe with OpenVINO tool.
Author: qhanson
Author-email: yifeihuang17@gmail.com
License: UNKNOWN
Platform: UNKNOWN
Classifier: Programming Language :: Python :: 3
Classifier: License :: OSI Approved :: MIT License
Classifier: Operating System :: OS Independent
Requires-Python: >=3.6
Description-Content-Type: text/markdown
License-File: LICENSE
Requires-Dist: openvino (>=2022.1.0)
Requires-Dist: opencv-python

# MediapipeOpenvino

This is a subset reimplemention of Mediapipe with OpenVINO tool.

# Usage Example

## Face Detection
``` Python

import cv2
import numpy as np
from MediapipeOpenvino.face_detection import FaceDetection

frame = cv2.imread("xx") ## Currently, we only support HWC format.
fd = FaceDetection()

re = fd(frame)

for batch in re: ## This version only support one image. This is a dummy loop.
    for dec in batch:
        img = cv2.rectangle(frame, (int(dec.xmin*w),int(dec.ymin*h)), (int(dec.xmax*w),int(dec.ymax*h)), (255, 0, 0), 1)
cv2.imwrite("test.jpg",frame)
```


# Copyright
All the copyright in this code lies on the [Mediapipe](https://github.com/google/mediapipe). We just follow the original implementation and use Python to achieve the same result.

