app/services/segmentation/detect.py

import argparse
import time
from pathlib import Path
import cv2
import torch
import torch.backends.cudnn as cudnn
from numpy import random
from PIL import Image

from models.experimental import attempt_load
from utils.datasets import LoadStreams, LoadImages
from utils.general import check_img_size, non_max_suppression, scale_coords, strip_optimizer, set_logging, increment_path
from utils.plots import plot_one_box
from utils.torch_utils import select_device, time_synchronized

def process_and_save_image(im0, xyxy, save_dir, filename, target_size=(224, 224)):
    # Crop and pad the image to make it square
    x_min, y_min, x_max, y_max = int(xyxy[0]), int(xyxy[1]), int(xyxy[2]), int(xyxy[3])
    width, height = x_max - x_min, y_max - y_min
    max_side = max(width, height)
    center_x, center_y = (x_min + x_max) // 2, (y_min + y_max) // 2

    new_x_min = max(0, center_x - max_side // 2)
    new_y_min = max(0, center_y - max_side // 2)
    new_x_max = min(im0.shape[1], new_x_min + max_side)
    new_y_max = min(im0.shape[0], new_y_min + max_side)

    cat_image = im0[new_y_min:new_y_max, new_x_min:new_x_max]
    padded_image = cv2.copyMakeBorder(
        cat_image,
        top=max(0, new_y_min - y_min),
        bottom=max(0, y_max - new_y_max),
        left=max(0, new_x_min - x_min),
        right=max(0, x_max - new_x_max),
        borderType=cv2.BORDER_CONSTANT, value=[0, 0, 0]
    )

    # Resize to 224x224 and convert to RGB
    resized_image = cv2.resize(padded_image, target_size)
    image_rgb = cv2.cvtColor(resized_image, cv2.COLOR_BGR2RGB)
    pil_image = Image.fromarray(image_rgb)
    output_path = save_dir / f"{filename}_cat.png"
    pil_image.save(output_path, 'PNG')
    print(f"Saved processed image to {output_path}")

def detect(save_img=False):
    source, weights, view_img, save_txt, imgsz = opt.source, opt.weights, opt.view_img, opt.save_txt, opt.img_size
    webcam = source.isnumeric() or source.endswith('.txt') or source.lower().startswith(
        ('rtsp://', 'rtmp://', 'http://'))

    # Directories
    save_dir = Path(increment_path(Path(opt.project) / opt.name, exist_ok=opt.exist_ok))  # increment run
    save_dir.mkdir(parents=True, exist_ok=True)  # make dir

    # Initialize
    set_logging()
    device = select_device(opt.device)
    half = device.type != 'cpu'  # half precision only supported on CUDA

    # Load model
    model = attempt_load(weights, map_location=device)  # load FP32 model
    imgsz = check_img_size(imgsz, s=model.stride.max())  # check img_size
    if half:
        model.half()  # to FP16

    # Set Dataloader
    dataset = LoadStreams(source, img_size=imgsz) if webcam else LoadImages(source, img_size=imgsz)
    
    # Run inference
    t0 = time.time()
    img = torch.zeros((1, 3, imgsz, imgsz), device=device)  # init img
    _ = model(img.half() if half else img) if device.type != 'cpu' else None  # run once

    for path, img, im0s, vid_cap in dataset:
        img = torch.from_numpy(img).to(device)
        img = img.half() if half else img.float()  # uint8 to fp16/32
        img /= 255.0  # 0 - 255 to 0.0 - 1.0
        if img.ndimension() == 3:
            img = img.unsqueeze(0)

        # Inference
        t1 = time_synchronized()
        pred = model(img, augment=opt.augment)[0]

        # Apply NMS
        pred = non_max_suppression(pred, opt.conf_thres, opt.iou_thres, classes=opt.classes, agnostic=opt.agnostic_nms)
        t2 = time_synchronized()

        # Process detections
        for i, det in enumerate(pred):  # detections per image
            p, im0 = (path[i], im0s[i].copy()) if webcam else (path, im0s)

            # Rescale boxes from img_size to im0 size
            if len(det):
                det[:, :4] = scale_coords(img.shape[2:], det[:, :4], im0.shape).round()

                # Iterate over detected objects
                for *xyxy, conf, cls in reversed(det):
                    if opt.save_txt:  # Save detection details if needed
                        pass

                    # Check if detected object is a cat (you can set your class filter to cats only)
                    if int(cls) in [14, 15, 16, 17, 18, 19, 20]:  # assuming 15 is the class index for 'cat'
                        filename = Path(p).stem
                        process_and_save_image(im0, xyxy, save_dir, filename)

    print(f'Done. ({time.time() - t0:.3f}s)')