about project:
video classification using the UCF101 and kinetics-700 action recognition dataset.
dataset_links
https://www.deepmind.com/open-source/kinetics
https://www.tensorflow.org/datasets/catalog/ucf101
['ApplyEyeMakeup','ApplyLipstick','Archery','BabyCrawling','BalanceBeam','BandMarching', 'BaseballPitch','BasketballDunk','Basketball','BenchPress','Biking', 'Billiards','BlowDryHair','BlowingCandles','BodyWeightSquats','Bowling', 'BoxingPunchingBag','BoxingSpeedBag','BreastStroke','BrushingTeeth']
- We used the 2 datasets to make a prediction but used ucf101 for production.
- We make many traials to get the final results.
- Deploy tensorflow model using flask.