Can't sucessfully transform the pdf coordinates into image coordinates.

[Andy718811]

Here is my code, the result is not very well.

import json
from PIL import Image, ImageDraw
import matplotlib.pyplot as plt
import fitz  # pip install pymupdf
import json
from PIL import Image, ImageDraw
import matplotlib.pyplot as plt

import json
from PIL import Image, ImageDraw
import fitz  # pip install pymupdf



import json
from PIL import Image, ImageDraw

class Chunk:
    def __init__(self, text, pos):
        self.text = text
        self.x1, self.x2, self.y1, self.y2 = pos

def load_chunks(json_path):
    with open(json_path, "r", encoding="utf-8") as f:
        data = json.load(f)
    return [Chunk(text=entry["text"], pos=entry["pos"]) for entry in data["chunks"]]

def get_table_bounds(chunks):
    x_min = min(c.x1 for c in chunks)
    x_max = max(c.x2 for c in chunks)
    y_min = min(c.y1 for c in chunks)
    y_max = max(c.y2 for c in chunks)
    return x_min, x_max, y_min, y_max

def pdf_chunk_to_image_bbox(chunk, x_min, x_max, y_min, y_max, img_w, img_h, padding=2):
    scale_x = (img_w-2) / (x_max - x_min)
    scale_y = (img_h-2) / (y_max - y_min)

    x1 = (min(chunk.x1,chunk.x2) - x_min) * scale_x 
    x2 = (max(chunk.x1,chunk.x2) - x_min) * scale_x
    y1 = (y_max - min(chunk.y2,chunk.y1)) * scale_y   # Y 軸翻轉
    y2 = (y_max - max(chunk.y2,chunk.y1)) * scale_y
    return min(x1, x2), min(y1, y2), max(x1, x2), max(y1, y2)


def visualize_chunks_on_cropped_image(json_path, image_path, box_color="red", box_width=2):
    # 載入圖片與標註
    image = Image.open(image_path).convert("RGB")
    img_w, img_h = image.size
    draw = ImageDraw.Draw(image)

    chunks = load_chunks(json_path)
    x_min, x_max, y_min, y_max = get_table_bounds(chunks)

    for c in chunks:
        x1, y1, x2, y2 = pdf_chunk_to_image_bbox(c, x_min, x_max, y_min, y_max, img_w, img_h)
        draw.rectangle([(x1, y1), (x2, y2)], outline=box_color, width=box_width)

    image.show()


if __name__ == "__main__":
    json_path=r"D:\下載\SciTSR\SciTSR\train\chunk\0707.0704v1.1.chunk"
    image_path=r"D:\下載\SciTSR\SciTSR\train\img\0707.0704v1.1.png"
    visualize_chunks_on_cropped_image(json_path, image_path)