C++ Demo - Facial Expression Recognition

[DaniAffCH]

This Pull Request introduces a C++ demo for utilizing and visualizing the Facial Expression Recognition (FER) model requested in issue #135 . The changes include:

• demo.cpp: This file contains the demo implementation, featuring a class YuNet for face detection (adapted from face_detection_yunet/demo.cpp to create a stand-alone demo) and a class FER for preprocessing and inference with the Facial Expression Recognition model.

• CMakeLists.txt: The related CMakeLists file necessary for building the program.

• README.md: The README file has been updated to include information about this new demo.

The demo offers two modes of inference: real-time using the camera or processing a static image.

The coding style in demo.cpp aligns with the conventions used in other demos and follows the structure of the corresponding demo.py. Notably, I opted to utilize the estimateAffine2D function from the OpenCV library instead of reimplementing a similar mechanism, as was done in the Python version.

Testing

The program has been tested both with a single person and with multiple people at the same time.
Here are some screenshots:

Single Person from the cam

Multi-person from static image

Build & Run

[fengyuentau]

Nice and clean 👍 We can proceed to merge after the following minor comments are solved.

[DaniAffCH]

Hi @fengyuentau thank you for your review. I just fixed the minor issues you pointed out.

[fengyuentau]

Oops, sorry. I thought it was std::format. Please ignore this comment.

________________________________ 发件人: Daniele Affinita ***@***.***> 发送时间: Monday, February 26, 2024 9:11:08 PM 收件人: opencv/opencv_zoo ***@***.***> 抄送: Yuantao Feng ***@***.***>; Mention ***@***.***> 主题: Re: [opencv/opencv_zoo] C++ Demo - Facial Expression Recognition (PR #233) @DaniAffCH commented on this pull request.

________________________________ In models/facial_expression_recognition/demo.cpp<#233 (comment)>:

+ Mat faces = faceDetectionModel.infer(frame);

+ vector<String> expressions; + + for (int i = 0; i < faces.rows; ++i) + { + Mat face = faces.row(i); + String exp = expressionRecognitionModel.infer(frame, face); + expressions.push_back(exp); + + int x1 = static_cast<int>(faces.at<float>(i, 0)); + int y1 = static_cast<int>(faces.at<float>(i, 1)); + int w = static_cast<int>(faces.at<float>(i, 2)); + int h = static_cast<int>(faces.at<float>(i, 3)); + float conf = faces.at<float>(i, 14); + + std::cout << cv::format("%d: x1=%d, y1=%d, w=%d, h=%d, conf=%.4f expression=%s\n", i, x1, y1, w, h, conf, exp.c_str()); Do you mean the format? ― Reply to this email directly, view it on GitHub<#233 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AEDL63UZ6YD3UDPTRUPTGS3YVSCWZAVCNFSM6AAAAABDZGVPJOVHI2DSMVQWIX3LMV43YUDVNRWFEZLROVSXG5CSMV3GSZLXHMYTSMBQHAZTSMBZGI>. You are receiving this because you were mentioned.Message ID: ***@***.***>

[fengyuentau]

👍 Thank you for the contribution!

[ShiqiYu]

@DaniAffCH Thank you for the contribution.