| import openvino as ov |
| import cv2 |
| import numpy as np |
|
|
| core = ov.Core() |
| model = core.read_model(model='models/horizontal-text-detection-0001.xml') |
| compiled_model = core.compile_model(model = model, device_name="CPU") |
| input_layer = compiled_model.input(0) |
| output_layer = compiled_model.output(0) |
|
|
| def preprocess_data(image, input_layer): |
| N, C, H, W = input_layer.shape |
| resized_image = cv2.resize(image, (W, H)) |
| input_image = np.expand_dims(resized_image.transpose(2, 0, 1), 0) |
| return input_image, resized_image |
|
|
| def predict_image(image, conf_threshold): |
| input_image, resized_image = preprocess_data(image, input_layer) |
| output_key = compiled_model.output("boxes") |
| boxes = compiled_model([input_image])[output_key] |
| |
| boxes = boxes[~np.all(boxes == 0, axis=1)] |
| return boxes, resized_image |
|
|
| def convert_result_to_image(bgr_image, resized_image, boxes, threshold=0.3, conf_labels=True): |
| |
| colors = {"red": (255, 0, 0), "green": (0, 255, 0)} |
| |
| (real_y, real_x), (resized_y, resized_x) = ( |
| bgr_image.shape[:2], |
| resized_image.shape[:2], |
| ) |
| ratio_x, ratio_y = real_x / resized_x, real_y / resized_y |
|
|
| |
| rgb_image = cv2.cvtColor(bgr_image, cv2.COLOR_BGR2RGB) |
|
|
| |
| for box in boxes: |
| |
| conf = box[-1] |
| if conf > threshold: |
| |
| |
| |
| (x_min, y_min, x_max, y_max) = [ |
| (int(max(corner_position * ratio_y, 10)) if idx % 2 else int(corner_position * ratio_x)) for idx, corner_position in enumerate(box[:-1]) |
| ] |
|
|
| |
| rgb_image = cv2.rectangle(rgb_image, (x_min, y_min), (x_max, y_max), colors["green"], 3) |
|
|
| |
| |
| if conf_labels: |
| rgb_image = cv2.putText( |
| rgb_image, |
| f"{conf:.2f}", |
| (x_min, y_min - 10), |
| cv2.FONT_HERSHEY_SIMPLEX, |
| 0.8, |
| colors["red"], |
| 1, |
| cv2.LINE_AA, |
| ) |
|
|
| return rgb_image |