digit_recogniser.py

import numpy as np
import cv2
import mnist_test
import keras.backend.tensorflow_backend as tfback
import tensorflow as tf

#code is taken from here : https://github.com/keras-team/keras/issues/13684#issuecomment-595054461

def _get_available_gpus():
    """Get a list of available gpu devices (formatted as strings).

    # Returns
        A list of available GPU devices.
    """
    #global _LOCAL_DEVICES
    if tfback._LOCAL_DEVICES is None:
        devices = tf.config.list_logical_devices()
        tfback._LOCAL_DEVICES = [x.name for x in devices]
    return [x for x in tfback._LOCAL_DEVICES if 'device:gpu' in x.lower()]
#experimental_list_devices is deprecated in tf 2.1


def get_img_contour_thresh(img):
    x, y, w, h = 0, 0, 300, 300
    gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
    blur = cv2.GaussianBlur(gray, (35, 35), 0)
    ret, thresh1 = cv2.threshold(blur, 70, 255, cv2.THRESH_BINARY_INV + cv2.THRESH_OTSU)
    thresh1 = thresh1[y:y + h, x:x + w]
    contours, hierarchy = cv2.findContours(thresh1, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE)[-2:]
    return img, contours, thresh1

def main():
    loaded_model = mnist_test.model()

    cap = cv2.VideoCapture(0)
    while (cap.isOpened()):
        ret, img = cap.read()
        img, contours, thresh = get_img_contour_thresh(img)
        ans1 = ''

        if len(contours) > 0:
            contour = max(contours, key=cv2.contourArea)
            if cv2.contourArea(contour) > 2500:
                x, y, w, h = cv2.boundingRect(contour)
                newImage = thresh[y:y + h, x:x + w]
                newImage = cv2.resize(newImage, (28, 28))
                newImage = np.array(newImage)
                newImage = newImage.flatten()
                newImage = newImage.reshape(1, 1,28,28)
                ans1 = loaded_model.predict(newImage)
                ans1=ans1.tolist()
                ans1 = ans1[0].index(max(ans1[0]))

        x, y, w, h = 0, 0, 300, 300
        cv2.rectangle(img, (x, y), (x + w, y + h), (0, 255, 0), 2)

        cv2.putText(img, " Deep Network : " + str(ans1), (10, 330),
                    cv2.FONT_HERSHEY_SIMPLEX, 0.7, (0, 0, 255), 2)

        #change the window size to fit screen properly
        #img = cv2.resize(img, (1000, 600))
        cv2.imshow("Frame", img)
        cv2.imshow("Contours", thresh)
        if cv2.waitKey(1) & 0xFF == ord('q'):
            break

    cap.release()
    cv2.destroyAllWindows()

# tfback._get_available_gpus = _get_available_gpus
main()