Follow the instruction,which file to run first

create your own dataset from Generate dataset.ipynb file or take any dataset of images from kaggle or anywhere--I havent uploaded any dataset folder for images,because it consumes space.So made it as short as possible to upload in git. for creation of dataset,opencv can be used,multiple codes are available for this. Firstly,run splitteddataset which helps in splitting the images dataset into 3 file-train,test and validation. Following this train the model in in keras,CNN and select the best model that gives high accuracy.(NOTE: Before copying this file,Its better to have knowledge on CNN,stride,paddding,filters,channels,pooling layer) Than move to another file,and use opencv code here aswell for detecting video,and feed the best trained model here.