Katsuya Hyodo

There seems to be a bug in NCS 2's internal API. https://software.intel.com/en-us/forums/computer-vision/topic/804818 And, I solved the problem of low precision. There was a mistake in the logic of preprocessing and...

The C ++ program is the same as the Intel sample program. I have not modified anything except to customize USB camera switching. And the latest OpenVINO 2019 R1.0.1 is...

Thank you for pointing out. The problem you pointed out has been happening for quite some time. No solution has yet been found in the official forums.

@gcharko2018 In fact, there is little difference in implementation of C++ program and Python program. I just reimplemented the C++ program in Python. I am sorry, but I do not...

I fixed Python's demo program. Thanks, @lilohuang !! With the help of lilohuang, the accuracy has improved greatly.

>should we convert IR models for different resolutions? Yes. That's because OpenVINO's input resolution, batch size and number of channels are fixed. And you need to fix the resolution 416x416...

The performance of OpenVINO depends on the total number of CPU cores. Please see the benchmark article below. I can not try because I do not have the following high-performance...

If you quantize a model, it may work even faster. **https://www.intel.ai/introducing-int8-quantization-for-fast-cpu-inference-using-openvino/#gs.c21087**

Please implement it yourself referring to the following program. There are a few lines that need to be changed. **https://github.com/PINTO0309/MobileNet-SSD-RealSense/blob/master/MultiStickSSDwithPiCamera_OpenVINO_NCS2.py** or **https://github.com/PINTO0309/OpenVINO-EmotionRecognition/blob/master/main.py**

1. Install Tensorflow v1.12.0 ```bash $ sudo -H pip3 install tensorflow==1.12.0 --upgrade ``` 2. Execute below. ```bash $ python3 /opt/intel/openvino/deployment_tools/model_optimizer/mo_tf.py \ --input_model /data/train/frozen_darknet_yolov3_tiny_model.pb \ --tensorflow_use_custom_operations_config /opt/intel/openvino/deployment_tools/model_optimizer/extensions/front/tf/yolo_v3_tiny.json \ --data_type FP16 \...