I am using TensorFlow object detection API to detect humans in video frames. I want to give the detection coordinates to the centroid tracker, to assign an ID for every detected object and avoid detecting them every frame as a new object.
This is the code I am using for detection and getting the bounding boxes
# Import packages
import os
import cv2
import numpy as np
import tensorflow as tf
import sys
# This is needed since the notebook is stored in the object_detection folder.
from pandas._libs import json
sys.path.append("..")
# Import utilites
from utils import label_map_util
from utils import visualization_utils as vis_util
# Name of the directory containing the object detection module we're using
MODEL_NAME = 'inference_graph'
VIDEO_NAME = 'test.mp4'
# direction for out put file
FILE_OUTPUT = 'C:/Mohammad/tensorflow1/models/research/object_detection/savedframes/out.avi'
# Grab path to current working directory
CWD_PATH = os.getcwd()
# Path to frozen detection graph .pb file, which contains the model that is used
# for object detection.
PATH_TO_CKPT = os.path.join(CWD_PATH,MODEL_NAME,'frozen_inference_graph.pb')
# Path to label map file
PATH_TO_LABELS = os.path.join(CWD_PATH,'training','label_map.pbtxt')
# Path to video
PATH_TO_VIDEO = os.path.join(CWD_PATH,VIDEO_NAME)
# Number of classes the object detector can identify
NUM_CLASSES = 2
# Load the label map.
# Label maps map indices to category names, so that when our convolution
# network predicts `5`, we know that this corresponds to `king`.
# Here we use internal utility functions, but anything that returns a
# dictionary mapping integers to appropriate string labels would be fine
label_map = label_map_util.load_labelmap(PATH_TO_LABELS)
categories = label_map_util.convert_label_map_to_categories(label_map, max_num_classes=NUM_CLASSES, use_display_name=True)
category_index = label_map_util.create_category_index(categories)
# Load the Tensorflow model into memory.
detection_graph = tf.Graph()
with detection_graph.as_default():
od_graph_def = tf.GraphDef()
with tf.gfile.GFile(PATH_TO_CKPT, 'rb') as fid:
serialized_graph = fid.read()
od_graph_def.ParseFromString(serialized_graph)
tf.import_graph_def(od_graph_def, name='')
sess = tf.Session(graph=detection_graph)
# Define input and output tensors (i.e. data) for the object detection classifier
# Input tensor is the image
image_tensor = detection_graph.get_tensor_by_name('image_tensor:0')
# Output tensors are the detection boxes, scores, and classes
# Each box represents a part of the image where a particular object was detected
detection_boxes = detection_graph.get_tensor_by_name('detection_boxes:0')
# Each score represents level of confidence for each of the objects.
# The score is shown on the result image, together with the class label.
detection_scores = detection_graph.get_tensor_by_name('detection_scores:0')
detection_classes = detection_graph.get_tensor_by_name('detection_classes:0')
# Number of objects detected
num_detections = detection_graph.get_tensor_by_name('num_detections:0')
# Open video file
video = cv2.VideoCapture('C:/Mohammad/tensorflow1/models/research/object_detection/test.mp4')
frame_width = int(video.get(3))
frame_height = int(video.get(4))
out = cv2.VideoWriter(FILE_OUTPUT, cv2.VideoWriter_fourcc('M', 'J', 'P', 'G'),
10, (frame_width, frame_height))
frame_index = 0
while(video.isOpened()):
# Acquire frame and expand frame dimensions to have shape: [1, None, None, 3]
# i.e. a single-column array, where each item in the column has the pixel RGB value
ret, frame = video.read()
frame_expanded = np.expand_dims(frame, axis=0)
# Perform the actual detection by running the model with the image as input
(boxes, scores, classes, num) = sess.run(
[detection_boxes, detection_scores, detection_classes, num_detections],
feed_dict={image_tensor: frame_expanded})
# Draw the results of the detection (aka 'visulaize the results')
vis_util.visualize_boxes_and_labels_on_image_array(
frame,
frame_index,
np.squeeze(boxes),
np.squeeze(classes).astype(np.int32),
np.squeeze(scores),
category_index,
use_normalized_coordinates=True,
line_thickness=8,
min_score_thresh=0.80)
#frame_index +=1
if ret == True:
out.write(frame)
# writing coordinates
coordinates = vis_util.return_coordinates(
frame,
frame_index,
np.squeeze(boxes),
np.squeeze(classes).astype(np.int32),
np.squeeze(scores),
category_index,
use_normalized_coordinates=True,
line_thickness=8,
min_score_thresh=0.80)
for coordinate in coordinates:
(ymin, ymax, xmin, xmax, acc, classification) = coordinate
height = ymax - ymin
width = xmax -xmin
crop = frame[ymin:ymin+height, xmin:xmin+width]
path = 'C:/Mohammad/A_Mohammad laptop/Tensorflow From C Drive/tensorflow1/models/research/object_detection/savecroped'
cv2.imwrite(os.path.join(path,'crop%d.jpg' %frame_index), frame)
textfile = open('filename_string' + ".json", "a")
textfile.write(json.dumps(coordinates))
textfile.write("\n")
frame_index =frame_index+1
# txt file
#textfile = open('filename_string' + ".txt", "w")
#textfile.write(str(coordinates))
#textfile.write("\n")
# All the results have been drawn on the frame, so it's time to display it.
cv2.imshow('Object detector', frame)
# Press 'q' to quit
if cv2.waitKey(1) == ord('q'):
break
# Clean up
video.release()
cv2.destroyAllWindows()
So the best way is to defining an empty list and storing all detection results (coordinates of bounding boxes) and use the update function from centroid class to track the objects. Also do not forget to empty the list after each detection, otherwise, you will get so many IDs for one object!