Taking data from video stream

[PradiptaPN]

Hi, first of all thank you for providing wonderful deep learning tools.

I am currently working on some projects that require me to classify image from video, such as face. I have trained the network based on MnistDemo example and saved it to json file so i can use it on main program.

my questions are:

1. how can i set image taken from video as input data?
2. Should i use MnistDemo classes such as circular buffer and dataset?

thank you for your support

[cbovar]

Hi,

1. How can I set image taken from video as input data?

a. Get frames from the video (as images)
b. Create volumes from images
c. Feed volumes to network and do the training / inference

2. Should I use MnistDemo classes such as circular buffer and dataset?

MnistDemo.DataSet has some specific part related to Mnist (MnistEntry) but you could create you own simple DataSet maybe like ConvNetSharp.Performance.Tests.Set:

    public class Set
    {
        public Volume<double>[] Inputs { get; set; }
        public Volume<double> Outputs { get; set; }
    }

CircularBuffer are just used to so some moving average of the testing and training accuracy. It's up to you to use it but it's not a must for your project.

[PradiptaPN]

thank you for your answer :D