BeaverDam dense annotation viewer

show_annotations.py

import cv2
import json
import pandas as pd
import numpy as np

def convert_to_pandas(content):
    events = []
    for obj in content:
        for f in obj['frames']:
            events.append({
                'fid' : f['frameid'],
                'oid' : obj['id'],
                'type' : obj['type'],
                'state' : f['state'],
                'x' : f['x'],
                'y' : f['y'],
                'w' : f['w'],
                'h' : f['h'],
            })
    
    df = pd.DataFrame(events)
    return df

if __name__ == '__main__':

    import argparse
    parser = argparse.ArgumentParser(description='Show dense annotations')
    parser.add_argument('-video', help='Input video file', required=True)
    parser.add_argument('-ann', help='Dense annotation file', required=True)
    args = parser.parse_args()

    with open(args.ann) as fh:    
        content = json.load(fh)

    df = convert_to_pandas(content)

    cap = cv2.VideoCapture(args.video)

    idx = 0
    while(True):
        ret, frame = cap.read()

        for _, obj in df[(df.fid == idx)].iterrows():
            xywh = np.squeeze(obj[['x', 'y', 'w', 'h']].values.astype(int))
            cv2.rectangle(frame, (xywh[0],xywh[1]), (xywh[0]+xywh[2],xywh[1]+xywh[3]), (255,0,0), 2, -1)
        
        # Display the resulting frame
        cv2.imshow('frame',frame)
        if cv2.waitKey(1) & 0xFF == ord('q'):
            break
        
        idx += 1