![]() My initial guess was that the way people draw is closely related to how people write. I used these 4 countries because these 4 countries had good number of images and they also do not share same alphabet/language. The stroke information contains 2 additional dimensions: typical imageįrom this input dataset, I collected image data of CAT, TIGER, LION, DOG for image recognition part of my project.įor country preiction part of my project I selected 4 countries: United States, BRASIL, RUSSIA and SOUTH KOREA. The drawing feature is a list of strokes and stroke is a list of X,Y and time (3 lists within a stroke) The dataset that google released contains images and several features related to image.įeatures include drawing_ID, category(what quickdraw asked to draw), timestamp, whether AI guessed correct or not, user's country and drawing.ĭrawing is represented as a list of list of list. how fast/slow did users draw their images.amount of information (details) exist within an image.The important features from XGboost model indicates that users' country can be identified based on Results of Country predictionįor Country prediction, models had lower accuracy than ones from image recognition. Therefore, they make quite different predictions. Since CNN model looks into pixels and XGBoost model looks into features that I calculated,įeatures are engineered differently for each model (meaning models analyze images completely differently). I made 4-way classifier prediction models for both image recognition and country prediction (Total of 4 models).įor image recognition, both CNN and XGBoost models had high prediction accuracy. To answer these questions, I prepared 2 prediction models Can machine learning models identify users' country based on their drawings? Can machine learning models distinguish similar drawings?Ģ. With this dataset, I wanted to answer following 2 questions:ġ. (Of course there is no prize for winning this game but it is super addicting!) If google AI predicts what user is drawing, user wins!.While user draws a picture, google AI will try to predict what user is drawing.user is asked to draw a picture of certain category in 20 seconds.The google quickdraw is an online pictionary game application where. Google Quickdraw released dataset that contains over 50 million drawings on. jpeg images used in this readme markdown file.note that there is no data stored in this repo. Jupyter notebook that runs python codes above.has python codes to run XGboost ensemble method algorithm for image recognitions and country prediction.has python codes to run CNN for image recognitions and country prediction.has python codes to set up json raw file into data that could be used for CNN and ensemble methods.This is my repository for the quick draw prediction model project ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |