Upload 9 files

.gitattributes

@@ -33,3 +33,7 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+examples/Test20.jpg filter=lfs diff=lfs merge=lfs -text
+examples/Test21.jpg filter=lfs diff=lfs merge=lfs -text
+examples/Test22.jpg filter=lfs diff=lfs merge=lfs -text
+examples/Test23.jpg filter=lfs diff=lfs merge=lfs -text

app.py

@@ -0,0 +1,437 @@
+import os
+from pathlib import Path
+from typing import List, Union
+from PIL import Image
+import ezdxf.units
+import numpy as np
+import torch
+from torchvision import transforms
+from ultralytics import YOLOWorld, YOLO
+from ultralytics.engine.results import Results
+from ultralytics.utils.plotting import save_one_box
+from transformers import AutoModelForImageSegmentation
+import cv2
+import ezdxf
+import gradio as gr
+import gc
+from scalingtestupdated import calculate_scaling_factor
+from scipy.interpolate import splprep, splev
+from scipy.ndimage import gaussian_filter1d
+
+birefnet = AutoModelForImageSegmentation.from_pretrained(
+    "zhengpeng7/BiRefNet", trust_remote_code=True
+)
+
+device = "cpu"
+torch.set_float32_matmul_precision(["high", "highest"][0])
+
+birefnet.to(device)
+birefnet.eval()
+transform_image = transforms.Compose(
+    [
+        transforms.Resize((1024, 1024)),
+        transforms.ToTensor(),
+        transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225]),
+    ]
+)
+
+
+def yolo_detect(
+    image: Union[str, Path, int, Image.Image, list, tuple, np.ndarray, torch.Tensor],
+    classes: List[str],
+) -> np.ndarray:
+    drawer_detector = YOLOWorld("yolov8x-worldv2.pt")
+    drawer_detector.set_classes(classes)
+    results: List[Results] = drawer_detector.predict(image)
+    boxes = []
+    for result in results:
+        boxes.append(
+            save_one_box(result.cpu().boxes.xyxy, im=result.orig_img, save=False)
+        )
+
+    del drawer_detector
+
+    return boxes[0]
+
+
+def remove_bg(image: np.ndarray) -> np.ndarray:
+    image = Image.fromarray(image)
+    input_images = transform_image(image).unsqueeze(0).to("cpu")
+
+    # Prediction
+    with torch.no_grad():
+        preds = birefnet(input_images)[-1].sigmoid().cpu()
+    pred = preds[0].squeeze()
+
+    # Show Results
+    pred_pil: Image = transforms.ToPILImage()(pred)
+    print(pred_pil)
+    # Scale proportionally with max length to 1024 for faster showing
+    scale_ratio = 1024 / max(image.size)
+    scaled_size = (int(image.size[0] * scale_ratio), int(image.size[1] * scale_ratio))
+
+    return np.array(pred_pil.resize(scaled_size))
+
+
+def make_square(img: np.ndarray):
+    # Get dimensions
+    height, width = img.shape[:2]
+
+    # Find the larger dimension
+    max_dim = max(height, width)
+
+    # Calculate padding
+    pad_height = (max_dim - height) // 2
+    pad_width = (max_dim - width) // 2
+
+    # Handle odd dimensions
+    pad_height_extra = max_dim - height - 2 * pad_height
+    pad_width_extra = max_dim - width - 2 * pad_width
+
+    # Create padding with edge colors
+    if len(img.shape) == 3:  # Color image
+        # Pad the image
+        padded = np.pad(
+            img,
+            (
+                (pad_height, pad_height + pad_height_extra),
+                (pad_width, pad_width + pad_width_extra),
+                (0, 0),
+            ),
+            mode="edge",
+        )
+    else:  # Grayscale image
+        padded = np.pad(
+            img,
+            (
+                (pad_height, pad_height + pad_height_extra),
+                (pad_width, pad_width + pad_width_extra),
+            ),
+            mode="edge",
+        )
+
+    return padded
+
+
+def exclude_scaling_box(
+    image: np.ndarray,
+    bbox: np.ndarray,
+    orig_size: tuple,
+    processed_size: tuple,
+    expansion_factor: float = 1.5,
+) -> np.ndarray:
+    # Unpack the bounding box
+    x_min, y_min, x_max, y_max = map(int, bbox)
+
+    # Calculate scaling factors
+    scale_x = processed_size[1] / orig_size[1]  # Width scale
+    scale_y = processed_size[0] / orig_size[0]  # Height scale
+
+    # Adjust bounding box coordinates
+    x_min = int(x_min * scale_x)
+    x_max = int(x_max * scale_x)
+    y_min = int(y_min * scale_y)
+    y_max = int(y_max * scale_y)
+
+    # Calculate expanded box coordinates
+    box_width = x_max - x_min
+    box_height = y_max - y_min
+    expanded_x_min = max(0, int(x_min - (expansion_factor - 1) * box_width / 2))
+    expanded_x_max = min(
+        image.shape[1], int(x_max + (expansion_factor - 1) * box_width / 2)
+    )
+    expanded_y_min = max(0, int(y_min - (expansion_factor - 1) * box_height / 2))
+    expanded_y_max = min(
+        image.shape[0], int(y_max + (expansion_factor - 1) * box_height / 2)
+    )
+
+    # Black out the expanded region
+    image[expanded_y_min:expanded_y_max, expanded_x_min:expanded_x_max] = 0
+
+    return image
+
+
+def resample_contour(contour):
+    # Get all the parameters at the start:
+    num_points = 1000
+    smoothing_factor = 5
+    spline_degree = 3  # Typically k=3 for cubic spline
+
+    smoothed_x_sigma = 1
+    smoothed_y_sigma = 1
+
+    # Ensure contour has enough points
+    if len(contour) < spline_degree + 1:
+        raise ValueError(f"Contour must have at least {spline_degree + 1} points, but has {len(contour)} points.")
+
+    contour = contour[:, 0, :]
+
+    tck, _ = splprep([contour[:, 0], contour[:, 1]], s=smoothing_factor)
+    u = np.linspace(0, 1, num_points)
+    resampled_points = splev(u, tck)
+
+    smoothed_x = gaussian_filter1d(resampled_points[0], sigma=smoothed_x_sigma)
+    smoothed_y = gaussian_filter1d(resampled_points[1], sigma=smoothed_y_sigma)
+
+    return np.array([smoothed_x, smoothed_y]).T
+
+
+def save_dxf_spline(inflated_contours, scaling_factor, height):
+    degree = 3
+    closed = True
+
+    doc = ezdxf.new(units=0)
+    doc.units = ezdxf.units.IN
+    doc.header["$INSUNITS"] = ezdxf.units.IN
+
+    msp = doc.modelspace()
+
+    for contour in inflated_contours:
+        try:
+            resampled_contour = resample_contour(contour)
+            points = [
+                (x * scaling_factor, (height - y) * scaling_factor)
+                for x, y in resampled_contour
+            ]
+            if len(points) >= 3:
+                if np.linalg.norm(np.array(points[0]) - np.array(points[-1])) > 1e-2:
+                    points.append(points[0])
+
+                spline = msp.add_spline(points, degree=degree)
+                spline.closed = closed
+        except ValueError as e:
+            print(f"Skipping contour: {e}")
+
+    dxf_filepath = os.path.join("./outputs", "out.dxf")
+    doc.saveas(dxf_filepath)
+    return dxf_filepath
+
+
+def extract_outlines(binary_image: np.ndarray) -> np.ndarray:
+    """
+    Extracts and draws the outlines of masks from a binary image.
+    Args:
+        binary_image: Grayscale binary image where white represents masks and black is the background.
+    Returns:
+        Image with outlines drawn.
+    """
+    # Detect contours from the binary image
+    contours, _ = cv2.findContours(
+        binary_image, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_NONE
+    )
+
+    # smooth_contours_list = []
+    # for contour in contours:
+    #     smooth_contours_list.append(smooth_contours(contour))
+    # Create a blank image to draw contours
+    outline_image = np.zeros_like(binary_image)
+
+    # Draw the contours on the blank image
+    cv2.drawContours(
+        outline_image, contours, -1, (255), thickness=1
+    )  # White color for outlines
+
+    return cv2.bitwise_not(outline_image), contours
+
+
+def shrink_bbox(image: np.ndarray, shrink_factor: float):
+    """
+    Crops the central 80% of the image, maintaining proportions for non-square images.
+    Args:
+        image: Input image as a NumPy array.
+    Returns:
+        Cropped image as a NumPy array.
+    """
+    height, width = image.shape[:2]
+    center_x, center_y = width // 2, height // 2
+
+    # Calculate 80% dimensions
+    new_width = int(width * shrink_factor)
+    new_height = int(height * shrink_factor)
+
+    # Determine the top-left and bottom-right points for cropping
+    x1 = max(center_x - new_width // 2, 0)
+    y1 = max(center_y - new_height // 2, 0)
+    x2 = min(center_x + new_width // 2, width)
+    y2 = min(center_y + new_height // 2, height)
+
+    # Crop the image
+    cropped_image = image[y1:y2, x1:x2]
+    return cropped_image
+
+
+def to_dxf(contours):
+    doc = ezdxf.new()
+    msp = doc.modelspace()
+
+    for contour in contours:
+        points = [(point[0][0], point[0][1]) for point in contour]
+        msp.add_lwpolyline(points, close=True)  # Add a polyline for each contour
+
+    doc.saveas("./outputs/out.dxf")
+    return "./outputs/out.dxf"
+
+
+def smooth_contours(contour):
+    epsilon = 0.01 * cv2.arcLength(contour, True)  # Adjust factor (e.g., 0.01)
+    return cv2.approxPolyDP(contour, epsilon, True)
+
+
+def scale_image(image: np.ndarray, scale_factor: float) -> np.ndarray:
+    """
+    Resize image by scaling both width and height by the same factor.
+
+    Args:
+        image: Input numpy image
+        scale_factor: Factor to scale the image (e.g., 0.5 for half size, 2 for double size)
+
+    Returns:
+        np.ndarray: Resized image
+    """
+    if scale_factor <= 0:
+        raise ValueError("Scale factor must be positive")
+
+    current_height, current_width = image.shape[:2]
+
+    # Calculate new dimensions
+    new_width = int(current_width * scale_factor)
+    new_height = int(current_height * scale_factor)
+
+    # Choose interpolation method based on whether we're scaling up or down
+    interpolation = cv2.INTER_AREA if scale_factor < 1 else cv2.INTER_CUBIC
+
+    # Resize image
+    resized_image = cv2.resize(
+        image, (new_width, new_height), interpolation=interpolation
+    )
+
+    return resized_image
+
+
+def detect_reference_square(img) -> np.ndarray:
+    box_detector = YOLO("./last.pt")
+    res = box_detector.predict(img, conf=0.05)
+    del box_detector
+    return save_one_box(res[0].cpu().boxes.xyxy, res[0].orig_img, save=False), res[
+        0
+    ].cpu().boxes.xyxy[0]
+
+
+def resize_img(img: np.ndarray, resize_dim):
+    return np.array(Image.fromarray(img).resize(resize_dim))
+
+
+def predict(image, offset_inches):
+    try:
+        drawer_img = yolo_detect(image, ["box"])
+        shrunked_img = make_square(shrink_bbox(drawer_img, 0.8))
+    except:
+        raise gr.Error("Unable to DETECT DRAWER, please take another picture with different magnification level!")
+
+    # Detect the scaling reference square
+    try:
+        reference_obj_img, scaling_box_coords = detect_reference_square(shrunked_img)
+    except:
+        raise gr.Error("Unable to DETECT REFERENCE BOX, please take another picture with different magnification level!")
+
+    # reference_obj_img_scaled = shrink_bbox(reference_obj_img, 1.2)
+    # make the image sqaure so it does not effect the size of objects
+    reference_obj_img = make_square(reference_obj_img)
+    reference_square_mask = remove_bg(reference_obj_img)
+
+    # make the mask same size as org image
+    reference_square_mask = resize_img(
+        reference_square_mask, (reference_obj_img.shape[1], reference_obj_img.shape[0])
+    )
+
+    try:
+        scaling_factor = calculate_scaling_factor(
+            reference_image_path="./Reference_ScalingBox.jpg",
+            target_image=reference_square_mask,
+            feature_detector="ORB",
+        )
+    except ZeroDivisionError:
+        scaling_factor = None
+        print("Error calculating scaling factor: Division by zero")
+    except Exception as e:
+        scaling_factor = None
+        print(f"Error calculating scaling factor: {e}")
+
+    # Default to a scaling factor of 1.0 if calculation fails
+    if scaling_factor is None or scaling_factor == 0:
+        scaling_factor = 1.0
+        print("Using default scaling factor of 1.0 due to calculation error")
+
+    # Save original size before `remove_bg` processing
+    orig_size = shrunked_img.shape[:2]
+    # Generate foreground mask and save its size
+    objects_mask = remove_bg(shrunked_img)
+
+    processed_size = objects_mask.shape[:2]
+    # Exclude scaling box region from objects mask
+    objects_mask = exclude_scaling_box(
+        objects_mask,
+        scaling_box_coords,
+        orig_size,
+        processed_size,
+        expansion_factor=1.5,
+    )
+    objects_mask = resize_img(
+        objects_mask, (shrunked_img.shape[1], shrunked_img.shape[0])
+    )
+
+    # Ensure offset_inches is valid
+    if scaling_factor != 0:
+        offset_pixels = (offset_inches / scaling_factor) * 2 + 1
+    else:
+        offset_pixels = 1  # Default value in case of invalid scaling factor
+
+    dilated_mask = cv2.dilate(
+        objects_mask, np.ones((int(offset_pixels), int(offset_pixels)), np.uint8)
+    )
+
+    # Scale the object mask according to scaling factor
+    # objects_mask_scaled = scale_image(objects_mask, scaling_factor)
+    Image.fromarray(dilated_mask).save("./outputs/scaled_mask_new.jpg")
+    outlines, contours = extract_outlines(dilated_mask)
+    shrunked_img_contours = cv2.drawContours(
+        shrunked_img, contours, -1, (0, 0, 255), thickness=2
+    )
+    dxf = save_dxf_spline(contours, scaling_factor, processed_size[0])
+
+    return (
+        shrunked_img_contours,
+        outlines,
+        dxf,
+        dilated_mask,
+        scaling_factor,
+    )
+
+
+if __name__ == "__main__":
+    os.makedirs("./outputs", exist_ok=True)
+
+    ifer = gr.Interface(
+        fn=predict,
+        inputs=[
+            gr.Image(label="Input Image"),
+            gr.Number(label="Offset value for Mask(inches)", value=0.075),
+        ],
+        outputs=[
+            gr.Image(label="Ouput Image"),
+            gr.Image(label="Outlines of Objects"),
+            gr.File(label="DXF file"),
+            gr.Image(label="Mask"),
+            gr.Textbox(
+                label="Scaling Factor(mm)",
+                placeholder="Every pixel is equal to mentioned number in inches",
+            ),
+        ],
+        examples=[
+            ["./examples/Test20.jpg", 0.075],
+            ["./examples/Test21.jpg", 0.075],
+            ["./examples/Test22.jpg", 0.075],
+            ["./examples/Test23.jpg", 0.075],
+        ],
+    )
+    ifer.launch(share=True)

last.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8ecf93886616e47bcbd997c9149521eab864aea3c4fa9ff48a95ab23d8ecf51e
+size 6254691

requirements.txt

@@ -0,0 +1,7 @@
+transformers
+ultralytics==8.3.9
+ezdxf
+gradio
+kornia
+timm
+einops

scalingtestupdated.py

@@ -0,0 +1,180 @@
+import cv2
+import numpy as np
+import os
+import argparse
+from typing import Union
+from matplotlib import pyplot as plt
+
+
+class ScalingSquareDetector:
+    def __init__(self, feature_detector="ORB", debug=False):
+        """
+        Initialize the detector with the desired feature matching algorithm.
+        :param feature_detector: "ORB" or "SIFT" (default is "ORB").
+        :param debug: If True, saves intermediate images for debugging.
+        """
+        self.feature_detector = feature_detector
+        self.debug = debug
+        self.detector = self._initialize_detector()
+
+    def _initialize_detector(self):
+        """
+        Initialize the chosen feature detector.
+        :return: OpenCV detector object.
+        """
+        if self.feature_detector.upper() == "SIFT":
+            return cv2.SIFT_create()
+        elif self.feature_detector.upper() == "ORB":
+            return cv2.ORB_create()
+        else:
+            raise ValueError("Invalid feature detector. Choose 'ORB' or 'SIFT'.")
+
+    def find_scaling_square(
+        self, reference_image_path, target_image, known_size_mm, roi_margin=30
+    ):
+        """
+        Detect the scaling square in the target image based on the reference image.
+        :param reference_image_path: Path to the reference image of the square.
+        :param target_image_path: Path to the target image containing the square.
+        :param known_size_mm: Physical size of the square in millimeters.
+        :param roi_margin: Margin to expand the ROI around the detected square (in pixels).
+        :return: Scaling factor (mm per pixel).
+        """
+        
+        contours, _ = cv2.findContours(
+            target_image, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_NONE
+        )
+
+        if not contours:
+            raise ValueError("No contours found in the cropped ROI.")
+
+        # # Select the largest square-like contour
+        largest_square = None
+        largest_square_area = 0
+        for contour in contours:
+            x_c, y_c, w_c, h_c = cv2.boundingRect(contour)
+            aspect_ratio = w_c / float(h_c)
+            if 0.9 <= aspect_ratio <= 1.1:
+                peri = cv2.arcLength(contour, True)
+                approx = cv2.approxPolyDP(contour, 0.02 * peri, True)
+                if len(approx) == 4:
+                    area = cv2.contourArea(contour)
+                    if area > largest_square_area:
+                        largest_square = contour
+                        largest_square_area = area
+
+        # if largest_square is None:
+        #     raise ValueError("No square-like contour found in the ROI.")
+
+        # Draw the largest contour on the original image
+        target_image_color = cv2.cvtColor(target_image, cv2.COLOR_GRAY2BGR)
+        cv2.drawContours(
+            target_image_color, largest_square, -1, (255, 0, 0), 3
+        )
+
+        # if self.debug:
+        cv2.imwrite("largest_contour.jpg", target_image_color)
+
+        # Calculate the bounding rectangle of the largest contour
+        x, y, w, h = cv2.boundingRect(largest_square)
+        square_width_px = w
+        square_height_px = h
+
+        # Calculate the scaling factor
+        avg_square_size_px = (square_width_px + square_height_px) / 2
+        scaling_factor = 0.5 / avg_square_size_px  # mm per pixel
+
+        return scaling_factor #, square_height_px, square_width_px, roi_binary
+
+    def draw_debug_images(self, output_folder):
+        """
+        Save debug images if enabled.
+        :param output_folder: Directory to save debug images.
+        """
+        if self.debug:
+            if not os.path.exists(output_folder):
+                os.makedirs(output_folder)
+            debug_images = ["largest_contour.jpg"]
+            for img_name in debug_images:
+                if os.path.exists(img_name):
+                    os.rename(img_name, os.path.join(output_folder, img_name))
+
+
+def calculate_scaling_factor(
+    reference_image_path,
+    target_image,
+    known_square_size_mm=12.7,
+    feature_detector="ORB",
+    debug=False,
+    roi_margin=30,
+):
+    # Initialize detector
+    detector = ScalingSquareDetector(feature_detector=feature_detector, debug=debug)
+
+    # Find scaling square and calculate scaling factor
+    scaling_factor = detector.find_scaling_square(
+        reference_image_path=reference_image_path,
+        target_image=target_image,
+        known_size_mm=known_square_size_mm,
+        roi_margin=roi_margin,
+    )
+
+    # Save debug images
+    if debug:
+        detector.draw_debug_images("debug_outputs")
+
+    return scaling_factor
+
+
+# Example usage:
+if __name__ == "__main__":
+    import os
+    from PIL import Image
+    from ultralytics import YOLO
+    from app import yolo_detect, shrink_bbox
+    from ultralytics.utils.plotting import save_one_box
+
+    for idx, file in enumerate(os.listdir("./sample_images")):
+        img = np.array(Image.open(os.path.join("./sample_images", file)))
+        img = yolo_detect(img, ['box'])
+        model = YOLO("./last.pt")
+        res = model.predict(img, conf=0.6)
+        
+        box_img = save_one_box(res[0].cpu().boxes.xyxy, im=res[0].orig_img, save=False)
+        # img = shrink_bbox(box_img, 1.20)
+        cv2.imwrite(f"./outputs/{idx}_{file}", box_img)
+        
+        print("File: ",f"./outputs/{idx}_{file}")
+        try:
+
+            scaling_factor = calculate_scaling_factor(
+                reference_image_path="./Reference_ScalingBox.jpg",
+                target_image=box_img,
+                known_square_size_mm=12.7,
+                feature_detector="ORB",
+                debug=False,
+                roi_margin=90,
+            )
+            # cv2.imwrite(f"./outputs/{idx}_binary_{file}", roi_binary)
+            
+            # Square size in mm
+            # square_size_mm = 12.7
+
+            # # Compute the calculated scaling factors and compare
+            # calculated_scaling_factor = square_size_mm / height_px
+            # discrepancy = abs(calculated_scaling_factor - scaling_factor)
+            # import pprint
+            # pprint.pprint({
+            #     "height_px": height_px,
+            #     "width_px": width_px,
+            #     "given_scaling_factor": scaling_factor,
+            #     "calculated_scaling_factor": calculated_scaling_factor,
+            #     "discrepancy": discrepancy,
+            # })
+
+
+            print(f"Scaling Factor (mm per pixel): {scaling_factor:.6f}")
+        except Exception as e:
+            from traceback import print_exc
+            print(print_exc())
+            print(f"Error: {e}")