app.py

import os 
os.environ["TF_ENABLE_ONEDNN_OPTS"] = "0"
os.environ["KERAS_BACKEND"] = "jax"

import cv2
import numpy as np
from PIL import Image
import matplotlib.pyplot as plt
import keras 


def resize_for_inference(input_image_path):
    # Load the input image
    image = cv2.imread(input_image_path)

    # Convert the image to RGB format (for compatibility with GrabCut)
    image_rgb = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)

    # Initialize the mask for GrabCut
    mask = np.zeros(image.shape[:2], np.uint8)

    # Define the rectangle for the GrabCut algorithm
    height, width = image.shape[:2]
    rect = (10, 10, width - 20, height - 20)  # Slightly smaller than the full image

    # Allocate memory for the GrabCut algorithm
    bgd_model = np.zeros((1, 65), np.float64)
    fgd_model = np.zeros((1, 65), np.float64)

    # Apply the GrabCut algorithm
    cv2.grabCut(image_rgb, mask, rect, bgd_model, fgd_model, 5, cv2.GC_INIT_WITH_RECT)

    # Convert the mask to binary (foreground is white, background is black)
    binary_mask = np.where((mask == 2) | (mask == 0), 0, 255).astype('uint8')

    # Resize the binary mask to the desired shape (960x720)
    resized_mask = cv2.resize(binary_mask, (720, 960), interpolation=cv2.INTER_AREA)

    # Further resize the mask to target size (224x224)
    target_size = (224, 224)
    final_resized_mask = cv2.resize(resized_mask, target_size, interpolation=cv2.INTER_AREA)

    final_resized_mask = np.expand_dims(final_resized_mask, axis=-1)

    # Save the resized binary mask
    

    return final_resized_mask
# Call the function with the input image path
resized_image=resize_for_inference("download.jpg")
resized_image.shape





loaded_model = keras.saving.load_model("hf://datasciencesage/bodym_measurement_model")

single_image_expanded = np.expand_dims(resized_image, axis=0) 
predicted_vales=loaded_model.predict(single_image_expanded)[0]
columns = ['ankle', 'arm-length', 'bicep', 'calf', 'chest',
           'forearm', 'height', 'hip', 'leg-length', 'shoulder-breadth',
           'shoulder-to-crotch', 'thigh', 'waist', 'wrist']

for body_type,measurement in zip(columns,predicted_vales):
    print(body_type ," is  ",measurement)