index.md

Plan of Attack:

//Step 1
function tagPosiblity(var hints){
    return index_of_possiblities_from_db;
}

//Step 2
function mappingToDimensions(var index_of_possiblities_from_db){
    return dimensions;
}

//Step 3
function matchingDim(var dimensions){
    return option_possiblities;
}

//Step 4:
function checkNumber(var option_possiblities){
    if (CNN(haar_cascade(option_possiblities)) == ans_found){
        return answer
    }
    else{
        reload()
    }
}

Attack!

1. Download all images and create generic images

   • There is csv file in assets folder which contain links to multiple images.

     Generic image:

     

2. Extract information regarding images and Numbers

   • Create two dictionary one of tags in images and other of their dimension
   • Extract Number image

3. Train Haar-Cascade

   • Using Generic Images and Number Create your own dataset of images and label them.

   • Then use that dataset to train Haar Cascade

     

4. Train Convolutional-Neural-Network

   • Now once we can identify Number from images, time to make small C.N.N. and train it on output of haar_cascade.detect(dataset)

     

5. Write JavaScript Extension

   • Now we have haar_cascade values and cnn model, it is time to all put together and make it work.