Image processing Accuracy calculation

For $30 — Ends 27 Jul 2018

I have a fully trained convolution neural network model (coded in tensorflow) designed and trained for capturing the oil spilled areas in ocean images. It can semantically segment the input images into two categories: background and the oil.
However, I need to extend the code and retrain it with a new dataset, so I can measure the segmentation accuracy of the input images as threshold the data and calculate the accuracy of detection. So the hard part was finished
Basically, the project has three main parts.
1) Calculate the accuracy of the segmentation using a suitable metric like that (example attached) .
2) Visualize the accuracies and precision with a graph/plot." And the graphs should categorized (example attached)
3) Run the whole fully convolutional neural network model by an exe file to install (pip – opencv – numpy – pillow – scipy – matblotlib – tensorflow – and any other needed function or library ) to avoid any problems with any normal user.
a. Close the exe file for the user and handle the open source of the project to me which is able to be developed in the future and be saved as exe file again.
b. Explain in details what have been achieved in equations if possible - $30.0


Sign up to apply →

Posted on freelancer.com





Know a freelancer that can do this job?

Earn lifetime commission for every freelancer you refer, by becoming an affiliate. Find out more

Share this job: