Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

ASL demo internal PR #4

Open
wants to merge 181 commits into
base: develop
Choose a base branch
from
Open

ASL demo internal PR #4

wants to merge 181 commits into from

Conversation

mpashchenkov
Copy link
Owner

No description provided.

Roman Donchenko and others added 5 commits December 16, 2020 18:02
Merge OpenVINO toolkit 2021.2 content into master
Merge OpenVINO toolkit 2021.3 content into master
Merge OpenVINO toolkit 2021.4 content into master
@mpashchenkov
Copy link
Owner Author

@dmatveev

@aDanPin, if interest is exist.

@@ -0,0 +1,12 @@
# Copyright (C) 2018-2019 Intel Corporation
Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

-2019?

Copy link
Owner Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Fixed.

private:
double smoothingFactor;
double value;
};
Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Is it taken from the original demo and is this the way it is called in the original demo?

Copy link
Owner Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I get this from Gaze estimation. This will be removed, because OMZ get unified FPS counter soon.

Copy link
Owner Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Unified metrics have appeared. Will be removed.

}
};

} // namespace custom
Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I believe every sample have it now, is it worth putting it to some common place?

Copy link
Owner Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

All demos use this stream_source (IFD soon). It is good point put this to common cpp_gapi folder.

Copy link
Owner Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Comment on lines 16 to 22
struct TrackedObject {
cv::Rect rect;
float confidence;

int object_id;

size_t frame_idx;
Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Can this structure be just plain? E.g. default field initializers and no constructor?

Copy link
Owner Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Fixed.

Comment on lines 190 to 201
if (key == 27) break; //esc
if (key == 48) id = 0; // person id
else if (key == 49) id = 1;
else if (key == 50) id = 2;
else if (key == 51) id = 3;
else if (key == 52) id = 4;
else if (key == 53) id = 5;
else if (key == 54) id = 6;
else if (key == 55) id = 7;
else if (key == 56) id = 8;
else if (key == 57) id = 9;
else if (key == 13) out_label_number = -1; // reset last gesture
Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

magik numbers here?

Copy link
Owner Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Key codes for 0...9 buttons

Copy link
Owner Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Added comments.

Comment on lines 156 to 164
std::copy_n(matR.begin<float>(), height * width,
prepared_mat[0].begin<float>() + 0 * batch * height * width
+ step * height * width);
std::copy_n(matG.begin<float>(), height * width,
prepared_mat[0].begin<float>() + 1 * batch * height * width
+ step * height * width);
std::copy_n(matB.begin<float>(), height * width,
prepared_mat[0].begin<float>() + 2 * batch * height * width
+ step * height * width);
Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

What I'd expect here is

  1. The preprocessing part (crop, resize, covertTo, split) - happening only once per a new frame - if it is possble.
  2. Only copy_n happenning for all (past+new) preprocessed frames to form the resulting blob.

Comment on lines 1 to 17
// Copyright (C) 2021 Intel Corporation
// SPDX-License-Identifier: Apache-2.0
//

#include "exponential_averager.hpp"

ExponentialAverager::ExponentialAverager(double smoothingFactor, double initValue):
smoothingFactor(smoothingFactor), value(initValue) {
}

double ExponentialAverager::getAveragedValue() const {
return value;
}

void ExponentialAverager::updateValue(double newValue) {
value = smoothingFactor * newValue + (1. - smoothingFactor) * value;
}
Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This thing looks quite tiny, it is used everywhere or just in a single place? If the latter, I'd drop a separate file here and put it right to the place where it is used.

Copy link
Owner Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Will be removed.

throw std::logic_error("Gesture file doesn't open.");
}
return labels;
}
Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

no newline here

Copy link
Owner Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Added

cv::destroyWindow(storage_window_name_);
}
}
}
Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

no newline here either, also the alignment "hangs"

Copy link
Owner Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Added

@@ -327,6 +327,7 @@ std::shared_ptr<FaceRecognizer> processingFaceGallery(const cv::gapi::GNetPackag
int id = 0;
for (const auto& item : fn) {
std::string label = item.name();
std::cout << label << std::endl;
Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

is it required? Seems like an unstructured random output

Copy link
Owner Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

No. Removed.

@mpashchenkov mpashchenkov changed the base branch from develop to master September 20, 2021 07:25
@mpashchenkov mpashchenkov changed the base branch from master to develop September 20, 2021 07:25
Vladimir and others added 19 commits November 17, 2021 15:30
…vl/support_tr_0016

Text Recognition 0016
…rieval-update-link-to-gallery.txt

demos/image_retrieval: update link to gallery.txt
* Add inference adapter

* Do not refer directly to model_executor attributes

* Move Openvino interface from the demo and AsyncPipeline to Adapter, implement RemoteAdapter, add documentation to ModelAdapter Abstract class

* Add the __call__ method to ModelBase class

* Refactoring

* Update get_input_layers(), get_output_layers(), add Metadata structure, separate the model reading from the loading

* Update model wrappers for object_detection_demo, except YOLO wrappers

* Update Deblurring, Monodepth demos

* Update bert_named_entity_recognition, bert_question_answering demos

* Move logging to OpenvinoAdapter, add class Core

* Ngraph removed from YOLO

* Fix

* Update human_pose_estimation_demo + Openpose wrapper

* Update hpe_associative_embedding wrapper, bert_question_answering_embedding demo

* Fix in object_detection_demo

* Delete extra EOF

* Update yolo.py

* Remove int2float(), add list of openvino precisions

* Small fixes

* Suggestions

* Add ability to read network from byte buffer.

* Update demos/common/python/openvino/model_zoo/model_api/adapters/openvino_adapter.py

* Update demos/common/python/openvino/model_zoo/model_api/adapters/openvino_adapter.py

* Remove trailing whitespaces

Co-authored-by: eizamaliev <eduard.zamaliev@intel.com>
…g-strategy

Change branching strategy

Align the strategy with https://github.com/openvinotoolkit/openvino/wiki/Branches:
master - development branch
releases/202?/? - release branches
…-develop-master

CONTRIBUTING.md: develop->master
* update install section

* update openvino-dev installation

* add note about pip version
@dmatveev
Copy link

dmatveev commented Jun 1, 2023

@mpashchenkov can you please close this PR? I believe its no longer relevant?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging this pull request may close these issues.