Unable to Process Frame with Gvawatermark in Intel® DL Streamer 2021.4.2

1. Inited the environment:

   source /opt/intel/openvino_2021.4.752/bin/setupvars.sh
   source /opt/intel/openvino_2021.4.752/data_processing/dl_streamer/bin/setupvars.sh

2. Downloaded IR model from:

   xml: https://storage.openvinotoolkit.org/repositories/open_model_zoo/2021.4/models_bin/3/person-vehicle-bike-detection-2004/FP32/person-vehicle-bike-detection-2004.xml
   bin: https://storage.openvinotoolkit.org/repositories/open_model_zoo/2021.4/models_bin/3/person-vehicle-bike-detection-2004/FP32/person-vehicle-bike-detection-2004.bin

3. Ran pipeline Command:

   gst-launch-1.0 filesrc location=/dsp/gvaWatermarkIssue/PBV_360p_720p.3gp ! decodebin ! queue ! gvadetect model=./model/person-vehicle-bike-detection-2004.xml device=CPU threshold=0.5 inference-interval=5 ! gvatrack tracking-type=short-term-imageless ! gvawatermark ! videoconvert ! fakesink sync=true

4. Encountered error: from element /GstPipeline:pipeline0/GstGvaWatermark:gvawatermark0/GstGvaWatermarkImpl:gvawatermarkimpl: gvawatermark has failed to process frame.

Use Intel® Deep Learning Streamer (Intel® DL Streamer) 2022.3 onwards.