Build a website with Face-API and database

Good afternoon! I found several examples of using the Face-API technology to detect emotions from videos and online broadcasts from webcams. [login to view URL] [login to view URL] Now I need to improve it a little. Goal: To create an online site based on Face-API technologies for monitoring people's behavior based on video recordings. The program should have the following simple functionality (everything is shown in detail in the table attached to the order): 1. Ability to upload videos: After downloading the video, the program must process it and split it into several images by frames, saving them to the server depending on the specified interval (each second, 3 seconds, 5 seconds, 10 seconds, 20 seconds, 30 seconds, 1 minute). Add a category as a placemark. Add the original total number of people to the database (manually per lesson). For example, we know that 10 people were present. Next, programmatically detect the faces in these frames. 2. After detecting the people, the program should mark the status of the person at a specific time in the form of labels in the database. For an example of columns - please see the attached file. 3. After the program has made the marks, it should to save them in the database: Create a function to download from the database the table in xlsx format (or csv) by such criterias as: - According to the results of the analysis, - According to the results of the analysis by category, - According to the results of the analysis of all downloads. Create a user-friendly interface in the form of a line graph with the emotions of people in the frames, taking into account the fluctuations in the total number of people in the frames (also based on the results of the analysis of the uploaded video and the results of all previously uploaded videos by category - see the example in the file). 4. (OPTIONAL) After checking each frame, you need to add the correct / incorrect function to determine the labels, so that the program remembers the action and correctly sets the labels in the second case. That is, you need to keep the results and constantly improve based on machine learning. Attached an excel file with a demo of the required interface and a database table.