Breaking News: Grepper is joining You.com. Read the official announcement!

create a dataset for cnn tensorflow

_loganthx answered on May 6, 2023 Popularity 5/10 Helpfulness 3/10

answer create a dataset for cnn tensorflow

create a dataset for cnn tensorflow

Comment

'''Generate dataset: (Suppose .py file is inside Dataset folder
and Labels are folder's names: Example => C:\Users\Dataset\Cats\cat1.png)'''
import os, cv2 as cv, numpy as np, random

DIR = os.getcwd()
Labels = [i for i in os.listdir(DIR) if not('.') in i]
data, Cut, X, y, New_W, New_H = [], [], [], [], 50, 50

for Root, dir, files in os.walk(DIR):
	if Root!=DIR: Cut.append(len(os.listdir(Root))) # Normalize samples
      
for R, d, f in os.walk(DIR):
	if R!= DIR: # Load the images and save like: [image, label]
		lbl = [(Labels[id]) for id, _ in enumerate(Labels) if Labels[id] in R]
		Imgs = [(cv.imread(f'{R}\\{img}', 0)/255.0) for img in os.listdir(R)]
		for img in Imgs[:min(Cut)]: 
			data.append([cv.resize(img, (New_W, New_H)), lbl[0]])
		print(f'Num of imgs for {lbl} label ==> {len(Imgs[:min(Cut)])}')
random.shuffle(data)

for img, label in data: 
	X.append(img)
	y.append(label)
np.array(X).reshape(-1, New_W, New_H, 1)   #1 if gray; 3 if rgb
  
np.save(f"{DIR}\\X.npx", X) # Save file in DIR directory
np.save(f"{DIR}\\y.npx", y)

Popularity 5/10 Helpfulness 3/10 Language python

Source: Grepper

Tags: dataset image python tensorflow

Link to this answer
Share Copy Link

Contributed on May 06 2023

_loganthx

0 Answers Avg Quality 2/10

create a dataset for cnn tensorflow

Contents

More Related Answers

create a dataset for cnn tensorflow

Grepper

Documentation

Social

Legal

Contact

Oops, You will need to install Grepper and log-in to perform this action.