can not find some head file

[LebronJames0423]

hello，when i follow your Using Tutorials (https://bestofcpp.com/repo/iwatake2222-aiine_transform), after successfully compile in vs2017, and i open the build folder in vs2017, first it tell me that there is no tensorflow,and i download it from https://github.com/tensorflow/tensorflow,and put it in

,and then , after compile successfully again , it through another problem

,that is "Aiine Transform\aiine_transform\inference_helper\third_party\cmakes..\tensorflow\tensorflow/lite/interpreter_builder.h(26): fatal error C1083: 无法打开包括文件: “flatbuffers/flatbuffers.h”: No such file or directory".
can you tell me how to slove this? thank you very much.

[iwatake2222]

Probably you missed the following commands, or it failed. Please make sure the download process succeeds when you run download_dependencies.sh

cd inference_helper/third_party/tensorflow
chmod +x tensorflow/lite/tools/make/download_dependencies.sh
tensorflow/lite/tools/make/download_dependencies.sh

[LebronJames0423]

Probably you missed the following commands, or it failed. Please make sure the download process succeeds when you run download_dependencies.sh

cd inference_helper/third_party/tensorflow
chmod +x tensorflow/lite/tools/make/download_dependencies.sh
tensorflow/lite/tools/make/download_dependencies.sh

thank for your reply,
i run your project in win10, after follow your device, i find the "download_dependencies.sh" in mediapipe folder,and i put it in the following folder

,and run the "chmod +x tensorflow/lite/tools/make/download_dependencies.sh" in cmd

, and after do this , i compile it again in cmake

but the problem still there, is there anything wrong with my operations?
by the way, can you upload your compiled vs project in your github? thank you very much~

[LebronJames0423]

download_dependencies.sh

should i compile tensorflow lite in win10 and put it in your project before i compile your project？

[iwatake2222]

I don't think you follow the instruction.
Please read this section carefully.

https://github.com/iwatake2222/aiine_transform#download

[LebronJames0423]

I don't think you follow the instruction. Please read this section carefully.

https://github.com/iwatake2222/aiine_transform#download

i successfully compiled it ,and it works well in vs2017 in win10, thank you. But i have another problem : The face mesh module of mediapipe includes model inference and pre- and post-processing. The pre- and post-processing code is written in C++. Did you extracted the code or wrote it by yourself? thanks a lot.

[iwatake2222]

👍

I don't understand what you mean by extract , but I referred to https://github.com/google/mediapipe/blob/master/docs/solutions/face_mesh.md#output , then I wrote it in C++ by myself.

[LebronJames0423]

👍

I don't understand what you mean by extract , but I referred to https://github.com/google/mediapipe/blob/master/docs/solutions/face_mesh.md#output , then I wrote it in C++ by myself.

your work is really a cool job!
why i run your project:
i want to use the face mesh model of mediapipe. it really quick when inferencing ,but i do not want to use mediapipe framework and tensorflow lite engine, i want to use another engine ,mindspore. what i did last few days is that i followed the mediapipe framework

,to extract the code of pre and post processing the result of face_landmark.tflite model. that's what i mean "extract" . but the result keypoints is not Accurate. so i want to confirm why the results is not accurate, the inference results is not accurate or pre and post processing code is not right, so after i find your blog, i think i maybe helpful for me, and it do.
next i will see how you write this project.

[iwatake2222]

i do not want to use mediapipe framework and tensorflow lite engine,

Really understood!!
I prefer not using heavy frameworks, too.

Porting pre/post logic to your project might be challenging, but hope my code helps.
Good luck