part2.md

Part 2: Understanding the Rival!

Files: get_img_data.py

Extract information regarding images

There are two information we are going to use:

Tags:

Now all Generic Images are numbered so by looking at each image we can judge tags of it and repeating it for all the images will lead to you to a file like this

This assets/tags.txt file contain tags of all images in numerical order.

medal,gold,one,first
medal,silver,two,second
medal,copper,third,three
turtle,turtles,four,brown shell

Convert those variable into dictionary with tags as key

tagPossiblity = {}
for i in range(0,len(img_tags)): ##each line
    for tags in img_tags[i]:	##each tag in that line
        ##tagPossiblity[tags] = list of where that tag is mentioned
        try:
            tagPossiblity[tags].append(i)
        except:
            tagPossiblity[tags] = [i]

get_img_data.py

Dimensions:

Dimensions can be extracted by opencv-python just open images and then images.shape will return height, width, channels of image

Save that into a variable

getDimension = []
for i in range(0,len(os.listdir('gen_images'))): ##serial wise
    data = cv2.imread('gen_images/'+str(i)+'.jpg')
    h,w,_ = data.shape ##channel not requered therefor _
    getDimension.append([w,h])

get_img_data.py

finaly add these variable to a js file

 with open('assets/imageData.js','w') as f:
    f.write('let getDimension = '+str(getDimension))
    f.write('\nlet tagPossiblity = '+str(tagPossiblity))
f.close()

Numbers

Just Extract numbers in num_img folder using magic wand of Photoshop/GIMP or use my dataset.

Unzip num_img.zip for my dataset!

Previous Part

Next Part