Issue with BYTETracker initialization in live tracking script

Question

I’m facing a TypeError when trying to initialize the BYTETracker in my live tracking script using the ByteTrack library. The error specifically occurs in the __init__ method of the BYTETracker class.

Here’s the relevant part of my code:

trackers = [BYTETracker(ByteTrackArgument), BYTETracker(ByteTrackArgument), BYTETracker(ByteTrackArgument)]

And the error message I’m encountering is:

TypeError: unsupported operand type(s) for +: 'type' and 'float'

I’ve attempted to resolve it by creating an instance of ByteTrackArgument like this:

trackers = [BYTETracker(ByteTrackArgument()), BYTETracker(ByteTrackArgument()), BYTETracker(ByteTrackArgument())]

However, the issue persists. It’s worth noting that I’m utilizing FaceDetectorYN from OpenCV for face detection in my script.

Any insights on what might be causing this error, especially in conjunction with the use of FaceDetectorYN, and how to address it?

Additional Context:

I’m using the ByteTrack library for live tracking. link
The error occurs in the __init__ method of the BYTETracker class.
I’m passing an instance of ByteTrackArgument to the BYTETracker constructor.
Face detection is performed using FaceDetectorYN from OpenCV. link

here is my full code :

import cv2
from bytetracker import BYTETracker
import numpy as np

print("OpenCV version", cv2.__version__)

class ByteTrackArgument:
    track_thresh = 0.5
    track_buffer = 50
    match_thresh = 0.8
    aspect_ratio_thresh = 10.0
    min_box_area = 1.0
    mot20 = False

MIN_THRESHOLD = 0.5  # Adjust this threshold as needed

# Initialize ByteTrackArgument
byte_track_argument = ByteTrackArgument()

# Initialize BYTETracker
trackers = [BYTETracker(ByteTrackArgument()), BYTETracker(ByteTrackArgument()), BYTETracker(ByteTrackArgument())]
def start_webcam_tracking():
    cap = cv2.VideoCapture(0)  # Use 0 for default webcam, or provide the webcam URL

    if not cap.isOpened():
        print("Error: Could not open the camera.")
        return

    while True:
        ret, frame = cap.read()
        if not ret:
            print("Error: Could not read frame from the camera.")
            break

        # Face detection code
        detector = cv2.FaceDetectorYN.create(r"C:\Users\gratu\live tracker\face_detection_yunet_2023mar.onnx", "", (2200, 1200), score_threshold=MIN_THRESHOLD)
        img_W = int(frame.shape[1])
        img_H = int(frame.shape[0])
        detector.setInputSize((img_W, img_H))

        detections = detector.detect(frame)[1]

        if detections is not None:
            for detection in detections:
                x, y, width, height = map(int, detection[:4])
                cv2.rectangle(frame, (x, y), (x + width, y + height), (0, 255, 0), 2)

                # Update the tracker with face bounding boxes
                tracker.update(np.array([[x, y, x + width, y + height]]), [frame.shape[0], frame.shape[1]])

        # Get tracked results from BYTETracker
        online_targets = tracker.get_online_targets()

        if online_targets is not None:
            for target in online_targets:
                x, y, x2, y2 = target  # Modify this part based on BYTETracker's output format
                cv2.rectangle(frame, (x, y), (x2, y2), (255, 0, 0), 2)

        cv2.imshow('Webcam with Face Detection and Tracking', frame)

        if cv2.waitKey(1) & 0xFF == ord('q'):
            break

    cap.release()
    cv2.destroyAllWindows()

if __name__ == "__main__":
    start_webcam_tracking()

this was the tutorial that i tried to replicate, they have used yolox for person detection, I tried to do the same for face detection

link

Leave a Comment Cancel reply