Hello,
i’m on personal project and i’m trying to detect custom object with openCV
i have already use the detectmultiiscale function with a haarcascade model but results are just under my satisfication
So i’m trying to use tensorflow and with a code that i found on internet (and it works):
while(True):
t1 = cv2.getTickCount()
ret, frame = camera.read()
frame_rgb = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)
frame_expanded = np.expand_dims(frame_rgb, axis=0)
(boxes, scores, classes, num) = sess.run(
[detection_boxes, detection_scores, detection_classes, num_detections],
feed_dict={image_tensor: frame_expanded})
vis_util.visualize_boxes_and_labels_on_image_array(
frame,
np.squeeze(boxes),
np.squeeze(classes).astype(np.int32),
np.squeeze(scores),
category_index,
use_normalized_coordinates=True,
line_thickness=8,
min_score_thresh=0.85)
if cv2.waitKey(1) == ord('q'):
break
it works with a SSDLITE model now i have to create my own custom model
So my question is : How to train an SSDLITE model ? and is that the best way for object detection ?