In [1]:
!pip install -q tflite-model-maker
In [2]:
import os
import numpy as np
import tensorflow as tf
assert tf.__version__.startswith('2')
import matplotlib.pyplot as plt
import tflite_model_maker
from glob import glob
from PIL import Image, ImageOps
2021-11-03 21:22:16.282316: W tensorflow/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'libcudart.so.11.0'; dlerror: libcudart.so.11.0: cannot open shared object file: No such file or directory 2021-11-03 21:22:16.282347: I tensorflow/stream_executor/cuda/cudart_stub.cc:29] Ignore above cudart dlerror if you do not have a GPU set up on your machine.
In [3]:
def thumbnail(image=None, desired_size=None):
"""Resizes original image as close as possible to desired size.
Preserves aspect ratio of original image.
Does not modify the original image.
:Parameters:
----------
image : PIL.Image
Input Image for AI model detection.
desired_size : (width, height)
Size expected by the AI model.
:Returns:
-------
PIL.Image
Resized image fitting for the AI model input tensor.
"""
assert image
assert desired_size
thumb = image.copy()
w, h = desired_size
try:
# convert from numpy to native Python int type
# that PIL expects
if isinstance(w, np.generic):
w = w.item()
w = int(w)
h = h.item()
h = int(h)
thumb.thumbnail((w, h))
except Exception as e:
msg = (f"Exception in "
f"PIL.image.thumbnail(desired_size={desired_size}):"
f"type(width)={type(w)}, type(height)={type(h)}"
f"\n{e}"
)
raise RuntimeError(msg)
return thumb
def resize(image=None, desired_size=None):
"""Pad original image to exact size expected by input tensor.
Preserve aspect ratio to avoid confusing the AI model with
unnatural distortions. Pad the resulting image
with solid black color pixels to fill the desired size.
Do not modify the original image.
:Parameters:
----------
image : PIL.Image
Input Image sized to fit an input tensor but without padding.
Its possible that one size fits one tensor dimension exactly
but the other size is smaller than
the input tensor other dimension.
desired_size : (width, height)
Exact size expected by the AI model.
:Returns:
-------
PIL.Image
Resized image fitting exactly the AI model input tensor.
"""
assert image
assert desired_size
thumb = image.copy()
delta_w = desired_size[0] - thumb.size[0]
delta_h = desired_size[1] - thumb.size[1]
padding = (0, 0, delta_w, delta_h)
new_im = ImageOps.expand(thumb, padding)
assert new_im.size == desired_size
return new_im
In [4]:
interpreter = tf.lite.Interpreter(model_path="ai_models/tflite-model-maker-falldetect-model-best.tflite")
interpreter.allocate_tensors()
input_details = interpreter.get_input_details()
output_details = interpreter.get_output_details()
input_shape = input_details[0]['shape']
width, height = 224, 224
input_shape
Out[4]:
array([ 1, 224, 224, 3], dtype=int32)
In [5]:
def prediction(path_image):
template_image_src = Image.open(path_image)
_thumbnail = thumbnail(image=template_image_src, desired_size=(width, height))
img = resize(image=_thumbnail, desired_size=(width, height))
img = np.expand_dims(img,0)
interpreter.set_tensor(input_details[0]['index'], img)
interpreter.invoke()
output_data = interpreter.get_tensor(output_details[0]['index'])
pred = "fall" if np.argmax(output_data[0])==0 else "no-fall"
return pred
Prediction on Single Image¶
In [6]:
path_image = "Images/fall_img_8.png"
pred = prediction(path_image)
pred
Out[6]:
'fall'
Prediction on Multiple Image¶
In [7]:
def pred_for_dir(image_list):
image_list = os.listdir('Images/')
frame_to_display = len(image_list)
columns = 6
rows = (frame_to_display // columns) + 1
fig = plt.figure(figsize=(20, 28))
plt_cnt = 1
for i in image_list:
image = Image.open('Images/'+i)
pred = prediction('Images/'+i)
fig.add_subplot(rows, columns, plt_cnt)
plt.axis("off")
plt_title = pred
plt.title(plt_title)
plt.imshow(image)
plt_cnt += 1
In [8]:
pred_for_dir('Images/')
Prediction using URL¶
In [9]:
import requests
from io import BytesIO
In [10]:
def predict_from_url(url):
response = requests.get(url)
template_image_src = Image.open(BytesIO(response.content))
_thumbnail = thumbnail(image=template_image_src, desired_size=(width, height))
img = resize(image=_thumbnail, desired_size=(width, height))
img = np.expand_dims(img,0)
interpreter.set_tensor(input_details[0]['index'], img)
interpreter.invoke()
output_data = interpreter.get_tensor(output_details[0]['index'])
pred = "fall" if np.argmax(output_data[0])==0 else "no-fall"
plt.title(pred)
plt.imshow(template_image_src)
return pred
predict_from_url(input('Enter URL->'))
Enter URL->https://previews.123rf.com/images/ljupco/ljupco2010/ljupco201000052/156473069-mature-man-sitting-in-an-armchair-and-reading-a-newspaper-in-a-living-room-at-home.jpg
Out[10]:
'no-fall'
In [ ]:
# sample fall images
# https://www.verywellhealth.com/thmb/S5_3kxnfNnzraYArYYSfeLOA9Ws=/1777x1333/smart/filters:no_upscale()/GettyImages-1016885118-5c074c6946e0fb0001d37d79.jpg
# https://i.pinimg.com/originals/af/15/0d/af150de073839d91c26d91fe015c039b.jpg
# https://encrypted-tbn0.gstatic.com/images?q=tbn:ANd9GcRvkYg_rLNNgrhcsNENb8dF0HZZP72-NsvnAA&usqp=CAU
# https://image.shutterstock.com/image-photo/elder-woman-lying-on-floor-260nw-786236107.jpg
# https://i.pinimg.com/474x/f2/14/be/f214bec1273997e4873e7a29053ad30d--elderly-care-safety-tips.jpg
# https://image.shutterstock.com/image-photo/elderly-woman-falling-down-home-260nw-1542815507.jpg
# https://image.shutterstock.com/image-photo/asian-senior-man-falling-down-260nw-1736873057.jpg
# sample non-fall images
# https://encrypted-tbn0.gstatic.com/images?q=tbn:ANd9GcRVSW_-L0x7w1p4ho82ClVB6t-wd3hu0rbSGw&usqp=CAU
# https://encrypted-tbn0.gstatic.com/images?q=tbn:ANd9GcRPxPHJPYa4qC-C1q8ZiVbrN-lP60muQOQV0w&usqp=CAU
# https://encrypted-tbn0.gstatic.com/images?q=tbn:ANd9GcTKHmd6AXI-dsPyMVGC0x_9vktzNykdtgDj0g&usqp=CAU
# https://encrypted-tbn0.gstatic.com/images?q=tbn:ANd9GcSXczMrp_LOv12u_RW-Y7GBHZWwtcnwXax8qw&usqp=CAU
# https://thumbor.forbes.com/thumbor/fit-in/900x510/https://www.forbes.com/health/wp-content/uploads/2021/04/medical_alerts_medicare-getty_creative.jpg
# https://bestlifeonline.com/wp-content/uploads/sites/3/2018/06/stressed-out-older-man.jpg
In [ ]: