Abandoned Object Detection using Deep learning

I am working on Abandoned Object detection, I want to know how I apply deep learning on it. As the pretrained network doesn't have data related to it.

And if I discuss about the existing work related to it, I found that they used synthetic data to create the network.

My queries are:-

1. Is it valid to use synthetic data?
2. Secondly, we need to pre-process the frames from video before classification through deep learning (as I read in most of the research). So, I am doing this process when I am working with NN and ML. So, what is the significance of using DL.

Please suggest.

Answer 1

[Jalaj Gambhir]

Hi,

Many a times real life data is not available, for those cases synthetic data can be used to train your model. This is perfectly valid. However, it is advised you should test your model on real life data.

You can do the task using end-to-end DL networks such as RCNN from Deep Learning Toolbox MATLAB.

[Monika Mehta]

Thank you for the response.