Image Preprocessing

Objective

In this article, we’ll walk you through the entire process of using a pre-trained model, loading it using the dnn module, image preprocessing using the blobfromImage method in OpenCV, and then finally making predictions....

Download the model files and Install Dependencies

You can download the mobile_net_ssd model here: https://github.com/chuanqi305/MobileNet-SSD/...

Load the Model:

Since we’re using caffe model we’ll use the cv2.dnn.readNetFromCaffe module to load our model. You will need these two types of files to work with any pre-trained model using dnn module:...

Image Preprocessing

So we’ll define a few variables for image preprocessing....

Make Predictions using the model

Now that we have our input ready, we have to explicitly set it as input using the setInput() method and then pass it to our model and use the forward method to generate predictions....

Below is the complete Implementation

Python3 import cv2 import dlib from imutils import face_utils   img = cv2.imread('object (1).png')   #--------Model Path---------# proto_file = 'SSD_MobileNet_prototxt.txt' model_file = 'SSD_MobileNet.caffemodel'   #------Variables for the Model ---------# classNames = {0: 'background',               1: 'aeroplane', 2: 'bicycle',               3: 'bird', 4: 'boat',               5: 'bottle', 6: 'bus', 7: 'car',               8: 'cat', 9: 'chair',               10: 'cow', 11: 'diningtable',               12: 'dog', 13: 'horse',               14: 'motorbike', 15: 'person',               16: 'pottedplant',               17: 'sheep', 18: 'sofa',               19: 'train', 20: 'tvmonitor'}   input_shape = (300, 300) mean = (127.5, 127.5, 127.5) scale = 0.007843   #---------Load The Model--------# net = cv2.dnn.readNetFromCaffe(proto_file, model_file)   #------image preprocessing----# blob = cv2.dnn.blobFromImage(img,                              scalefactor=scale,                              size=input_shape,                              mean=mean,                              swapRB=True)  # since our image is already in the BGR form   net.setInput(blob) results = net.forward() for i in range(results.shape[2]):           # confidence     confidence = round(results[0, 0, i, 2],2)     if confidence > 0.7:                   # class id         id = int(results[0, 0, i, 1])                    # 3-6 contains the coordinate         x1, y1, x2, y2 = results[0, 0, i, 3:7]                    # print(x1,y1,x2,y2)         # scale these coordinates to out image pixel         ih, iw, ic = img.shape         x1, x2 = int(x1*iw), int(x2*iw)         y1, y2 = int(y1 * ih), int(y2 * ih)         cv2.rectangle(img,                       (x1, y1),                       (x2, y2),                       (0, 200, 0), 2)         cv2.putText(img, f'{classNames[id]}:{confidence*100}',                     (x1+30, y1-30),                     cv2.FONT_HERSHEY_DUPLEX,                     1, (255, 0, 0), 1)     # print(results[0,0,i,:])   img = cv2.resize(img, (640, 720)) cv2.imshow('Image', img) # cv2.imwrite('output1.jpg',img) # Uncomment this line to save the output cv2.waitKey()...

What next?

...