SegCloth.py

from transformers import pipeline
from PIL import Image
import numpy as np
from io import BytesIO
import base64

# Initialize segmentation pipeline
segmenter = pipeline(model="mattmdjaga/segformer_b2_clothes")

def encode_image_to_base64(image):
    buffered = BytesIO()
    image.save(buffered, format="PNG")
    return base64.b64encode(buffered.getvalue()).decode('utf-8')

def segment_clothing(img, clothes=["Hat", "Upper-clothes", "Skirt", "Pants", "Dress", "Belt", "Left-shoe", "Right-shoe", "Scarf"]):
    # Segment image
    segments = segmenter(img)

    # Create list of masks and their corresponding clothing types
    mask_list = []
    for s in segments:
        if s['label'] in clothes:
            mask_list.append((s['mask'], s['label']))

    result_images = []

    # Paste all masks on top of each other
    for mask, clothing_type in mask_list:
        current_mask = np.array(mask)
        final_mask_bis = Image.fromarray(current_mask)
        img.putalpha(final_mask_bis)
        imageBase64 = encode_image_to_base64(img)
        result_images.append((clothing_type, imageBase64))

    return result_images