仅在 cv2 视频帧中的 ROI 部分预测模型类答案

【问题标题】：Predict model classes only on ROI part in cv2 video frame仅在 cv2 视频帧中的 ROI 部分预测模型类
【发布时间】：2021-01-25 11:43:02
【问题描述】：

以下是我的代码，其中包含 ROI 矩形。我想仅根据 ROI 框中捕获的图像而不是整个图像来预测我的模型

import numpy as np
import cv2
from keras.models import load_model
import keras
model = keras.models.load_model('Desktop/BCD_nothing_rmsprop.h5')
def sketch_transform(frame):
    return frame

cam_capture = cv2.VideoCapture(0)
cv2.destroyAllWindows()
upper_left = (50, 50)
bottom_right = (300, 300)
while True:
    _, image_frame = cam_capture.read()
    #Rectangle marker
    r = cv2.rectangle(image_frame, upper_left, bottom_right, (100, 50, 200), 5)
    rect_img = image_frame[upper_left[1] : bottom_right[1], upper_left[0] : bottom_right[0]]
    
    sketcher_rect = rect_img
    sketcher_rect = sketch_transform(sketcher_rect)
    
    
    #Replacing the sketched image on Region of Interest
    image_frame[upper_left[1] : bottom_right[1], upper_left[0] : bottom_right[0]] = sketcher_rect
    classes = model.predict(np.array([rect_img]))
    cv2.imshow("Sketcher ROI", image_frame)
    k = cv2.waitKey(30) & 0xff
    if k ==27:
        break
        
cam_capture.release()
cv2.destroyAllWindows()

如何仅预测矩形 ROI 或变量 rect_img 内的图像。

【问题讨论】：

标签： image numpy opencv image-processing keras

【解决方案1】：

之前创建resize函数。那就试试吧

while True:
    _, image_frame = cam_capture.read()
    #Rectangle marker
    r = cv2.rectangle(image_frame, upper_left, bottom_right, (100, 50, 200), 5)
    rect_img = image_frame[upper_left[1] : bottom_right[1], upper_left[0] : bottom_right[0]]

    sketcher_rect = rect_img
    ##Resize ROI first by create function
    sketcher_rect = resize_function(sketcher_rect,(width,height))


    #Replacing the sketched image on Region of Interest
    image_frame[upper_left[1] : bottom_right[1], upper_left[0] : bottom_right[0]] = sketcher_rect
    classes = model.predict(sketcher_rect.reshape(1,width,height,3))
    cv2.imshow("Sketcher ROI", image_frame)
    k = cv2.waitKey(30) & 0xff
    if k ==27:
        break
cam_capture.release()
cv2.destroyAllWindows()

【讨论】：