ligthedition.py

# -*- coding: utf-8 -*-
"""ligthEdition.ipynb

Automatically generated by Colaboratory.

Original file is located at
    https://colab.research.google.com/drive/1y6lGi22s7FxqWt4uW7l1WmHJsp2YDh1f

# **Automatic Ligth ajust**

_________________________________________________________________________________________________________________________________________________________________________________________________________
Angel Toledo Flores<br>
Ingeniero Mecánico Agrícola/Ingeniero en Sistemas Computacionales.<br>
**E-mail**: contactoangeltoledo@gmail.com<br>
 [Linkedin profile](https://www.linkedin.com/in/angel-toledo/) | [Portafolio de proyectos](https://resume.ingesaurio.com) <br> 
 [ DataSet: kagle](https://www.kaggle.com/code/shruthimshruthim/google-scrapped-image)

**`Just put the path of the image inside the inference function and enjoy the magic.`**

## Functions
"""

import cv2
from matplotlib.colors import hsv_to_rgb
import numpy as np
import matplotlib.pyplot as plt

def read_img(PATH):
  """
  Read img and transform to HSV
  """
  img = cv2.imread(PATH)
  img = cv2.resize(img,(128,128))
  hsv = cv2.cvtColor(img,cv2.COLOR_BGR2HSV) / 255.0
  return hsv

def inferencia(image_path):
  """
  Obtein inferencie of the model
  """
  image_hsv = read_img(image_path)
  images = []
  images.append(image_hsv[:,:,2])
  images.append(image_hsv[:,:,2])

  image_v = np.array(images)
  np.shape(image_hsv)
  predict_gama = model.predict(image_v)
  VV = image_v[0]
  print(np.shape(image_hsv))
  print(np.shape(VV))
  VV = VV.reshape((128,128))
  image_hsv[:,:,2] = VV / predict_gama[0]
  RGB_PREDICT = hsv_to_rgb(image_hsv)
  plt.imshow(RGB_PREDICT)
  plt.show()

"""## Import Model"""

from keras.models import load_model

!git clone https://github.com/toledoangel/automatic-image-brightness-adjustment.git

model = load_model('/content/automatic-image-brightness-adjustment/models/convPerceptronModel.h5')

"""## Inference"""

inferencia('/content/1002640661_c6adc3335c_m.jpg')