SalehAhmad's picture
new
ee83b99
raw
history blame
3.47 kB
import os
os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2'
os.environ['CUDA_VISIBLE_DEVICES'] = '0'
import tensorflow as tf
import tf_bodypix
from tf_bodypix.api import download_model, load_model, BodyPixModelPaths
from tf_bodypix.draw import draw_poses
from tensorflow.keras import preprocessing
import cv2
import json
from matplotlib import pyplot as plt
import numpy as np
from calculations import measure_body_sizes
import gradio as gr
import pandas as pd
# Load BodyPix model
bodypix_model = load_model(download_model(BodyPixModelPaths.MOBILENET_FLOAT_50_STRIDE_16))
rainbow = [
[110, 64, 170], [143, 61, 178], [178, 60, 178], [210, 62, 167],
[238, 67, 149], [255, 78, 125], [255, 94, 99], [255, 115, 75],
[255, 140, 56], [239, 167, 47], [217, 194, 49], [194, 219, 64],
[175, 240, 91], [135, 245, 87], [96, 247, 96], [64, 243, 115],
[40, 234, 141], [28, 219, 169], [26, 199, 194], [33, 176, 213],
[47, 150, 224], [65, 125, 224], [84, 101, 214], [99, 81, 195]
]
def process_images(front_img, side_img, real_height_cm):
fimage_array = preprocessing.image.img_to_array(front_img)
simage_array = preprocessing.image.img_to_array(side_img)
# bodypix prediction
frontresult = bodypix_model.predict_single(fimage_array)
sideresult = bodypix_model.predict_single(simage_array)
front_mask = frontresult.get_mask(threshold=0.75)
side_mask = sideresult.get_mask(threshold=0.75)
# preprocessing.image.save_img(f'{output_path}/frontbodypix-mask.jpg',front_mask)
# preprocessing.image.save_img(f'{output_path}/sidebodypix-mask.jpg',side_mask)
front_colored_mask = frontresult.get_colored_part_mask(front_mask, rainbow)
side_colored_mask = sideresult.get_colored_part_mask(side_mask, rainbow)
# preprocessing.image.save_img(f'{output_path}/frontbodypix-colored-mask.jpg',front_colored_mask)
# preprocessing.image.save_img(f'{output_path}/sidebodypix-colored-mask.jpg',side_colored_mask)
frontposes = frontresult.get_poses()
front_image_with_poses = draw_poses(
fimage_array.copy(), # create a copy to ensure we are not modifing the source image
frontposes,
keypoints_color=(255, 100, 100),
skeleton_color=(100, 100, 255)
)
sideposes = sideresult.get_poses()
side_image_with_poses = draw_poses(
simage_array.copy(), # create a copy to ensure we are not modifing the source image
sideposes,
keypoints_color=(255, 100, 100),
skeleton_color=(100, 100, 255)
)
# print(np.array(simage).shape)
# print(np.array(side_colored_mask).shape)
# preprocessing.image.save_img(f'{output_path}/frontbodypix-poses.jpg', front_image_with_poses)
# preprocessing.image.save_img(f'{output_path}/sidebodypix-poses.jpg', side_image_with_poses)
body_sizes = measure_body_sizes(side_colored_mask, front_colored_mask, sideposes, frontposes, real_height_cm, rainbow)
measurements_df = pd.DataFrame([body_sizes[0]])
return measurements_df
# Create the Gradio interface
interface = gr.Interface(
fn=process_images,
inputs=[
gr.Image(label="Upload Front Pose"),
gr.Image(label="Upload Side Pose"),
gr.Number(label="Enter Height (cm)")
],
outputs=[
gr.DataFrame(label="Body Measurements")
],
title="Body Sizing System Demo",
description="Upload two images: Front View and Side View, and input the height in cm."
)
# Launch the app
interface.launch(share=False)