Information about data collection

[Dntfreitas]

Hey, @mrayinteractive,

My colleagues and I have been working on a scientific research project focused on the QuickDraw dataset. We've been conducting some experiments. However, in order to ensure the accuracy of our findings, we need to gather more information about the data collection process that took place prior to the model training and the development of the game.

Specifically, we're curious about a few things:

• During the sketching process, were participants given a time limit to complete their drawings or was the time unlimited? For example, we noticed that one sketch of an envelope appears to be unfinished. We're interested in understanding the reasons behind this:
  
• We're wondering if the current data is being fed back into the model and used to update the recognized key. It would be great to know more about this aspect.
• Lastly, we're wondering if it would be possible for us to access the dataset that was used to train the model. Having access to this data would greatly contribute to our research.

Any additional information you can provide would be greatly appreciated! Thanks in advance for your help.

[alihaydar29]

Hi, the dataset consists of sketch recognation, the time given for the drawings is about 10 seconds. The drawings are simple sketches. So there are many errors in the drawings. But since the number of data is large, it does not seriously reduce the success of the model. By the way, there is also text in the drawings (ex : horse). You have to split the data(test-train). It is actually quite simple. You cannot process all the data because the data is too big. You have to take a part of it.
My english is not very good, I hope you understand :)