Turorial for image (graphs) recognition

Hi everyone,

I would like to ask you about recommendation of tutorial for image recognition.
In my case the images are graphs (scatter plot), where I have dataset for signal and background events, i.e. 2 segmentation classification.

Thank you very much in advance.

Hi @Sjuzn, Please refer to this gist to perform image recognition using tensorflow. As you have 2 classes please change the neurons in the last dense layer to 2. Thank You.

1 Like

I will try it! Thanks a lot @Kiran_Sai_Ramineni

Hi @Kiran_Sai_Ramineni,

is it problem for scatter plots that in case of CNN they need to be converted into images and these images have resolution e.g. 256x256. A lot of information from the scatter plots will diisappear, right?
How can be this problem solved?

Hi @Sjuzn, may i know which format your scatter plots are ? Thank You.

My scatter plots can be converted to png (1812x961 pixels).
Originally, scatter plots are on canvases coming from the ROOT CERN software.
However, if my images 1812x961 are “downgraded” to 256x256 pixels in CNN, then it is not possible for CNN to “see” points in scatter plot, right?