Week 24 - Final results

Modeled images results

Raw images results

Conclusions

In general, three important points are concluded:

• Raw images are more complicated for networks than modeled ones
• Recurrence introduces an improvement in the prediction.
• A more complex network provides better results in more complex motions.

You can find more details about the project in the memory: https://github.com/RoboticsLabURJC/2017-tfm-nuria-oyaga/blob/master/Documentation/Memoria-TFM/TFM_NURIA_OYAGA.pdf