app.py

import gradio as gr

#import sys
#import os
import pandas as pd
import numpy as np
#import cv2
#import matplotlib.pyplot as plt
#from PIL import Image
#import keras
#import tensorflow as tf
#from keras.models import Model
#from keras.optimizers import Adam
#from keras.applications.vgg16 import VGG16, preprocess_input
#from keras.applications.vgg19 import VGG19, preprocess_input
#from keras.preprocessing.image import ImageDataGenerator
#from keras.callbacks import ModelCheckpoint, EarlyStopping
#from keras.layers import Dense, Dropout, Flatten, MaxPooling2D, Conv2D
#from pathlib import Path
#from sklearn.metrics import accuracy_score


from keras.models import model_from_json
#from keras.preprocessing import image
#from keras.applications.vgg16 import VGG16, preprocess_input
#import heapq

file = open("focusondriving.json", 'r')
model_json2 = file.read()
file.close()
loaded_model = model_from_json(model_json2)
#loaded_model.load_weights("focusondriving.h5")

class_dict = {
    'c0': 'hands on the wheel',
    'c1': 'mobile in right hand',
    'c2': 'talking on the phone with right hand',
    'c3': "mobile in left hand",
    'c4': 'talking on the phone with left hand',
    'c5': 'touching at the dash',
    'c6': 'drinking',
    'c7': 'reaching behind',
    'c8': 'touching the head',
    'c9': 'looking to the side'
}

def predict_image(pic):
   # img = image.load_img(pic, target_size=(224, 224))
   # x = image.img_to_array(img) 
   # x = np.expand_dims(x, axis=0)
   # x = preprocess_input(x)
  #  preds = loaded_model.predict(x)
   # preds = list(preds[0])

    #list_desc_order = heapq.nlargest(2, range(len(preds)), key=preds.__getitem__)
    #result1 = f'c{list_desc_order[0]}'
    #result2 = '-'
    #result2_ = 0
    #if preds[list_desc_order[1]] > 0.3:
     #   result2 = f'c{list_desc_order[1]}'
      #  result2_  = round(preds[list_desc_order[1]], 2)
    #txt = f"category {directory} result 1 {result1} {round(preds[list_desc_order[0]],2)} | result2 {result2} {result2_}"
    #txt = f"categoria {directory}"
    
    #score = round(preds[list_desc_order[0]], 2)*100
    #score = int(score)
    #txt2 = f"resultado: {class_dict.get(result1)}     probabilidad {score}%"
    txt3="pepe"
    return txt3
    
    
iface = gr.Interface(
    predict_image,
    [
        
        gr.inputs.Image(source="upload",type="filepath", label="Imagen")
    ],

    "text",
    
    
    
    interpretation="default",
    title = 'FER - Facial Expression Recognition',
    description = 'Probablemente nos daremos cuenta de que muchas veces se miente cuando se tratan las emociones, ¿pero nuestra cara también miente? https://saturdays.ai/2022/03/16/detectando-emociones-mediante-imagenes-con-inteligencia-artificial/ ',
    
    theme = 'grass'
 )


   
iface.launch()