Revision history [back]

Thanks to the OpenCV developers group, we are now able to import Mask-RCNN's mask_rcnn_inception_v2_coco_2018_01_28 model in OpenCV. But I was not able to import either mask_rcnn_resnet101_atrous_coco_2018_01_28 or mask_rcnn_inception_resnet_v2_atrous_coco_2018_01_28, while trying to improve the Mask-RCNN's results with better backbone architectures. I am getting the following error:

cv2.error: OpenCV(3.4.3) /Users/snayak/opencv/opencv-3.4.3/modules/dnn/src/tensorflow/tf_importer.cpp:1008: error: (-215:Assertion failed) values.type() == CV_32FC1 in function 'populateNet'

The .pbtxt file used above was generated using the following command:

python3 ~/opencv/opencv-3.4.3/samples/dnn/tf_text_graph_mask_rcnn.py --input=mask_rcnn_resnet101_atrous_coco_2018_01_28/frozen_inference_graph.pb --output=mask_rcnn_resnet101_atrous_coco_2018_01_28.pbtxt

Thanks to the OpenCV developers group, we are now able to import Mask-RCNN's mask_rcnn_inception_v2_coco_2018_01_28 model in OpenCV. But I was not able to import either mask_rcnn_resnet101_atrous_coco_2018_01_28 or mask_rcnn_inception_resnet_v2_atrous_coco_2018_01_28, while trying to improve the Mask-RCNN's results with better backbone architectures. These models are from the Tensorflow Detection Model Zoo

I am getting the following error:

cv2.error: OpenCV(3.4.3) /Users/snayak/opencv/opencv-3.4.3/modules/dnn/src/tensorflow/tf_importer.cpp:1008: error: (-215:Assertion failed) values.type() == CV_32FC1 in function 'populateNet'

The .pbtxt file used above was generated using the following command:

python3 ~/opencv/opencv-3.4.3/samples/dnn/tf_text_graph_mask_rcnn.py --input=mask_rcnn_resnet101_atrous_coco_2018_01_28/frozen_inference_graph.pb --output=mask_rcnn_resnet101_atrous_coco_2018_01_28.pbtxt

Thanks to the OpenCV developers group, we are now able to import Mask-RCNN's mask_rcnn_inception_v2_coco_2018_01_28 model in OpenCV. But I was not able to import either mask_rcnn_resnet101_atrous_coco_2018_01_28 or mask_rcnn_inception_resnet_v2_atrous_coco_2018_01_28, while trying to improve the Mask-RCNN's results with better backbone architectures. These models are from the Tensorflow Detection Model Zoo

I am getting the following error:

cv2.error: OpenCV(3.4.3) /Users/snayak/opencv/opencv-3.4.3/modules/dnn/src/tensorflow/tf_importer.cpp:1008: error: (-215:Assertion failed) values.type() == CV_32FC1 in function 'populateNet'

The .pbtxt file used above was generated using the following command:

python3 ~/opencv/opencv-3.4.3/samples/dnn/tf_text_graph_mask_rcnn.py --input=mask_rcnn_resnet101_atrous_coco_2018_01_28/frozen_inference_graph.pb --output=mask_rcnn_resnet101_atrous_coco_2018_01_28.pbtxt

Thanks to the OpenCV developers group, we are now able to import Mask-RCNN's mask_rcnn_inception_v2_coco_2018_01_28 model in OpenCV. But I was not able to import either mask_rcnn_resnet101_atrous_coco_2018_01_28 or mask_rcnn_inception_resnet_v2_atrous_coco_2018_01_28, while trying to improve the Mask-RCNN's results with better backbone architectures. These models are from the Tensorflow Detection Model Zoo

I am getting the following error:

The .pbtxt file used above was generated using the following command:

python3 ~/opencv/opencv-3.4.3/samples/dnn/tf_text_graph_mask_rcnn.py --input=mask_rcnn_resnet101_atrous_coco_2018_01_28/frozen_inference_graph.pb --output=mask_rcnn_resnet101_atrous_coco_2018_01_28.pbtxt
--output=mask_rcnn_resnet101_atrous_coco_2018_01_28.pbtxt