Final Report

[cleebp]

Heres a list of what still needs work (besides general editing):

• 3D CNN Implementation subsection under methodology section, this subsection will be where we walk through in relevant detail what our code is doing, what preprocess.py does and why, and what it produces to be then used by conv_nn.py and what it does with that and why.
• Could use some more discussion in the Results and Discussions section on our results, comparing them to other Kaggle submissions. Could also talk about the meaning of our accuracy as discussed in Section 5: Concluding Remarks
• Finally if anyone understands 3D CNNs (i hardly do) the subsection on them in the Related Works section could use maybe one more little paragraph on what exactly they do or how they are different from normal neural networks.

[cleebp]

One thing that might be really good to make would be an overview figure diagram which demonstrates what the 3D CNN is doing with the stage1/ dataset and its output, if we made a really good one we could include it in our report and reference it in our methodology section, and it could also make up a large part of our poster (the bigger the resolution on the figure the better so when we print it out on the poster it won't look blurry).