Example: 03-Machine-Learning/00-TensorFlow/blazepalm_detection.py

# 本作品采用MIT许可证授权。
# Copyright (c) 2013-2025 OpenMV LLC. All rights reserved.
# https://github.com/openmv/openmv/blob/master/LICENSE
#
# This example shows off Google's MediaPipe Palm Detection model.
#
# NOTE: This exaxmple requires an OpenMV Cam with an NPU like the AE3 or N6 to run real-time.

import csi
import time
import ml
from ml.postprocessing.mediapipe import BlazePalm

# Initialize the sensor.
csi0 = csi.CSI()
csi0.reset()
csi0.pixformat(csi.RGB565)
csi0.framesize(csi.VGA)

# BlazePalm requires a square image for the best results.
csi0.window((400, 400))

# Load built-in palm detection model
model = ml.Model("/rom/palm_detection_full_192.tflite", postprocess=BlazePalm(threshold=0.4))
print(model)

# Line connections between hand joints for drawing the hand skeleton.
palm_lines = ((0, 1), (1, 2), (2, 3), (3, 4), (4, 0), (0, 5), (5, 6))

clock = time.clock()
while True:
    clock.tick()
    img = csi0.snapshot()

    # palms is a list of ((x, y, w, h), score, keypoints) tuples
    for r, score, keypoints in model.predict([img]):
        ml.utils.draw_predictions(img, [r], ("palm",), ((0, 0, 255),), format=None)

        # keypoints is a ndarray of shape (7, 2)
        # 0 - wrist (x, y)
        # 1 - index finger mcp (x, y)
        # 2 - middle finger mcp (x, y)
        # 3 - ring finger mcp (x, y)
        # 4 - pinky mcp  (x, y)
        # 5 - thumb cmc (x, y)
        # 6 - thumb mcp (x, y)
        #
        # mcp = Metacarpophalangeal Joint - the knuckle
        # cmc = Carpometacarpal Joint - the base of the thumb
        ml.utils.draw_skeleton(img, keypoints, palm_lines, kp_color=(255, 0, 0), line_color=(0, 255, 0))

    print(clock.fps(), "fps")