OpenCV object detection run error
- OpenCV => 3.4
- Operating System - Ubuntu 16.04
- Python - 3.5
Following the instructions given in "TensorFlow Object Detection API" I'm trying to run my trained graph based on Inception v2 SSD in OpenCV, but I get the following errors: When I used not optimized for inference graph I get the following error:
[ INFO:0] Initialize OpenCL runtime...
OpenCV Error: Assertion failed (!_aspectRatios.empty()) in PriorBoxLayerImpl, file /io/opencv/modules/dnn/src/layers/prior_box_layer.cpp, line 207
Traceback (most recent call last):
File "opencvInf.py", line 9, in <module>
cvOut = cvNet.forward()
cv2.error: /io/opencv/modules/dnn/src/layers/prior_box_layer.cpp:207: error: (-215) !_aspectRatios.empty() in function PriorBoxLayerImpl
After that I tried to optimize the graph using optimize_for_inference.py script, but it does not solve the problem: Now I get another error;
OpenCV Error: Unspecified error (Input layer not found: FeatureExtractor/InceptionV2/InceptionV2/Mixed_4c/Branch_0/Conv2d_0a_1x1/BatchNorm/FusedBatchNorm) in connect, file /io/opencv/modules/dnn/src/tensorflow/tf_importer.cpp, line 553
Traceback (most recent call last):
File "opencvInf.py", line 3, in <module>
cvNet = cv.dnn.readNetFromTensorflow('opt_graph.pb', 'graph.pbtxt')
cv2.error: /io/opencv/modules/dnn/src/tensorflow/tf_importer.cpp:553: error: (-2) Input layer not found: FeatureExtractor/InceptionV2/InceptionV2/Mixed_4c/Branch_0/Conv2d_0a_1x1/BatchNorm/FusedBatchNorm in function connect
Could somebody tell how to solve this?
@Vahe, We need to know the difference between a public model and yours. What changes have been made in a config file? Moreover
OpenCV 3.4.1 or higher is required.(according to wiki page).The only change which I have done in original conf file during model training is that I have only 3 classes. I checked my OpenCV version is 3.4.0. Could it be the reason of this?