nhgowtham / asmdataset

Building CNN architecture for microstructures processing

asmDataset

Here is the master program for Carnegie Mellon University Material Science Department.

Language Using: Python3.5

Test&Learning

Floder Test&Learning is the current Testing & Learning methods which may be used in the future programming

Workflow

• Put the code floder under the same directory with the Micrograph floder

• Initialize.py to get cropped images and their labels

• TODO

Crop the graphs

Data initialization, cut the graphs into certain shape, e.g. $200\times 200$ pixels for these reasons:

• Using the valuable data for the further image representation. Remove the pixels that are useless, including scales, black edges, etc.

• The Micrographs usually have $>400$ pixels shape, by cropping the graphs, we could get more training data for the training system.

TODO

Questions:

is there any loss of transfering from rgb to gray?

TEST VERSION: To use the crop.py, run terminal in the data floder, run the crop.py file.