search

TomStog / Infrared-BP

The dataset used for the "Non-Contact Blood Pressure Estimation using infrared motion magnified facial video" publication. The code developed is to fit the data to the reference Blood Pressure values.
9 stars 0 forks source link

coordinate #3

Closed chengzhangsuper closed 10 months ago

chengzhangsuper commented 10 months ago

Can you share the volunteer RGB screenshot dataset?

TomStog commented 10 months ago

Unfortunately I can't due to privacy protocols applied.

TomStog commented 10 months ago

Basically I just recorded a few seconds with our camera in RGB mode before switching to Infrared

chengzhangsuper commented 10 months ago

Can you upload some pictures used in the code? I want to run it, thank you img = cv2.imread('18.png') image = cv2.imread('08.png') with open('18.txt', 'w') as file:

TomStog commented 10 months ago

Again, I'm very sorry but I can't, volunteers' data cannot be shared.

chengzhangsuper commented 10 months ago

I’m having a holiday these days so I forgot to reply. If you help me, I’m willing to share my future data.

TomStog commented 10 months ago

I uploaded a sample image for you to study the experimental setup. The face has been swaped with an artificially generated one. Further instructions can be found in the 3rd section of our paper, called "EXPERIMENTS". I also twiked a bit the "face_tracking_dnn_photo.py", "hand_detect4bloodpressure_photo.py" files. Nothing too exciting, just some minor additions in order to save the coordinates of the face and hand in MATLAB's coordinate system format. If your future video data setup can match our setup, that would be a tremendous help.

TomStog commented 10 months ago

And please let me know when you're planning to start builiding your dataset. Thank you.

chengzhangsuper commented 10 months ago

My teacher asked me to prepare the data set after understanding the code and running it successfully.

TomStog commented 10 months ago

I'm willing to offer an intensive course (Skype/Zoom/Microsoft Teams) covering the methodology, the literature I utilized, and a thorough explanation of the code to assist you in successfully reproducing my results, if you're able to secure a commitment from at least 30 volunteers and maintain the video format closely resembling the uploaded sample image and provide the data within a month.

TomStog commented 9 months ago

Check this: https://www.mdpi.com/2673-7426/4/1/25