ValueError: cannot reshape array

[Tsiga__George]

I am using openvino_2020.1.033 for my exercises. I am currently trying to run the sample text_recognition_in_python_demo. 

I have downloaded the sample models as specified. eg. python downloader.py --name text-spotting-0001-recognizer-decoder.

When I try to run the program using 

python text_spotting_demo.py -m_m "models/text-spotting-0001-detector.xml" -m_te "models/text-spotting-0001-recognizer-encoder.xml" -m_td "models/text-spotting-0001-recognizer-decoder.xml" -i sign.jpg --keep_aspect_ratio

I get the following error

[ INFO ] Creating Inference Engine...
[ INFO ] Loading network files:
        models/text-spotting-0001-detector.xml
        models/text-spotting-0001-detector.bin
[ INFO ] Loading network files:
        models/text-spotting-0001-recognizer-encoder.xml
        models/text-spotting-0001-recognizer-encoder.bin
[ INFO ] Loading network files:
        models/text-spotting-0001-recognizer-decoder.xml
        models/text-spotting-0001-recognizer-decoder.bin
[ INFO ] Loading IR to the plugin...
MFX: Unsupported extension: sign.jpg
[ERROR:0] global C:\jenkins\workspace\OpenCV\OpenVINO\2019R4\build\windows\opencv\modules\videoio\src\cap.cpp (116) cv::VideoCapture::open VIDEOIO(CV_IMAGES): raised OpenCV exception:

OpenCV(4.2.0-openvino) C:\jenkins\workspace\OpenCV\OpenVINO\2019R4\build\windows\opencv\modules\videoio\src\cap_images.cpp:253: error: (-5:Bad argument) CAP_IMAGES: can't find starting number (in the name of file): sign.jpg in function 'cv::icvExtractPattern'

[ ERROR ] Failed to open "sign.jpg"
[ INFO ] Starting inference...
To close the application, press 'CTRL+C' here or switch to the output window and press ESC key

I am not sure If I am doing anything wrong.
My Input Image (sign.jpg)

My Folder Structure

[Max_L_Intel]

Hi George.

You're doing this correctly. We were able to reproduce this issue on Windows, so we filled a bug to developers team to get this fixed. 
However, I can say there's no such error on Linux using the same demo. So this bug is specific to Windows.

Thank you for reporting this.

Best regards, Max.

[Anjaneya_Srujit_Ram]

Hi,

I got the same issue (MFX: Unsupported extension) with cv2.VideoCapture() on windows machine with OpenVINO 2020 R1. Can we expect the bug to be resolved in the next release? 

Thank you and Regards,

Anjaneya Srujit.

[Max_L_Intel]

Hi, Ramachandruni, Anjaneya Srujit.

Please let me know what video file format do you try to process with the text spotting demo? I've tested a couple of MPEG4 h.264 videos on Windows machine, and although I also see this "unsupported extension" warning message, it seems to be working correctly.

Unfortunately we can't provide comments on future releases, but as I mentioned, we reported this image files bug to developers team so they are aware of it. 

Thanks.
Best regards, Max.

[Anjaneya_Srujit_Ram]

Hi Max,

I am processing the MP4 file, which is converted from the YUV file using FFMPEG with libx264 codec. The issue is resolved by downloading and adding "opencv_videoio_ffmpeg420_64.dll"  in "opencv\bin" directory. Now, I am not getting any errors or warnings regarding (MFX: Unsupported Extension).

Thank you and Regards,

Anjaneya Srujit.

[Max_L_Intel]

Hi Anjaneya Srujit.

We are glad that you were able to solve this issue. Thanks very much for sharing your findings with the OpenVINO community.

Best regards, Max.

[Max_L_Intel]

Hi George.

We've heard back from developers team that on Windows platform there FFmpeg backend should be additionally downloaded (it's not installed by default) to handle single image.

FFmpeg backend is optional on Windows (OpenCV community FFmpeg plugin). There's downloader script from the package: openvino\opencv\ffmpeg-download.ps1 (right click - Run with PowerShell).

Thanks.
Best regards, Max.