kingpeter2015

I wrote one test: ``` auto calc_mediasoup = [](uint64_t ts){ return static_cast((ts & 0x1FFFFFC0) / 64); }; auto calc_mediasoup_modified = [](uint64_t ts){ return static_cast((ts & 0x3FFFFFC0) / 64); }; auto...

@jmillan There are int, int16_t, int32_t, int64_t etc types in C/C++, however, there is no 24-bit signed integer. And using the lowest 24 bits of a *uint32_t* type to assign...

you can check the dependency with following tools. [Dependencies.zip](https://github.com/kingpeter2015/libovmatting/files/8315144/Dependencies.zip)

Please build 'ovmatting' first. 

Maybe you should downgrade your openvino version to 2021.2.185

@972461099 AFA, libovmatting library only supports two matting method: METHOD_BACKGROUND_MATTING_V2 and METHOD_MODNET. It is defined in 'include/ovmatter.h'. METHOD_BACKGROUND_MATTING_V2 method uses model defined in 'pytorch_mobilenetv2.bin/pytorch_mobilenetv2.xml'. METHOD_MODNET method uses model defined in...

@972461099 The attachment file is origin onnx file i've used in the library. Please Ref [modnet.zip](https://github.com/kingpeter2015/libovmatting/files/6530223/modnet.zip)

@1234269 some opset node is define in latest version of pytorch which openvino can not parse. So there are only two ways to solve the problem: 1 You can train...

@1234269 InstanceNorm2d is define in latest version of pytorch which openvino can not parse. So there are only two ways to solve the problem: 1 You can train your model...

@wenjunyoung The file pytorch_xxx.pth is trained with some new version of pytorch which openvino can not parse. We have met these troubles during development, and modified something. I've forgotten the...