How to assign different color for different classes in object detection model?

Greetings, I have two classes, which are labelled as normal and cancer cells. How to draw bounding boxes with different colors that referred to each class?

newp2 = imread([pathname filename]);
[bboxes,scores,labels] = detect(detector,newp2,'Threshold',0.7);
annotations = string(labels) + ": " + string(scores);
I = insertObjectAnnotation(newp2,'rectangle',bboxes,cellstr(annotations));
figure
imshow(I)

Answer 1

[Sachin]

Hi

I understand that you want to assign the different color for different classes in object detection model.

To draw bounding boxes with different colors for each class, you can modify the ‘Color’ parameter based on the class label,

Here is an example:

newp2 = imread([pathname filename]);
[bboxes,scores,labels] = detect(detector,newp2,'Threshold',0.7);
annotations = string(labels) + ": " + string(scores);
%loop through each detected object
for i=1:numel(bboxes)
  label = labels(i);
  bbox = bboxes(i,:);
  score = scores(i);
  
  if label=='inormal'
     color = 'b'
  elseif label == 'cancer'
  
    color = 'r'
  else 
    color = 'g'

insertObjectAnnotation(newp2,'rectangle',bboxes,cellstr(annotations),'Color',color);
 end

Please refer the following page for more information about ‘insertobjectannotation’

https://www.mathworks.com/help/vision/ref/insertobjectannotation.html#namevaluepairarguments

[Fahmi Akmal Dzulkifli]

First of all, thank you Sir for the respond. When I execute the code, there is an error, where the index exceeds the number of array elements. For example, the bboxes is equal to (219x4 single), but the i is equal to 220. How can I correct it?

[Sachin]

Ok Then just try finding the number of rows in bboxes in this case 219. So replace the numel with number of rows

[Fahmi Akmal Dzulkifli]

I already tried it but the values for 'label' and 'labels are not same.The 'label' has (1x1 categorical) while the 'labels' has (219x1 categorical). Can you identify the error?

[filename, pathname] = uigetfile(('*.jpg;*.png;*.tiff;*.bmp,'));
newp2 = imread([pathname filename]);
[bboxes,scores,labels] = detect(detector,newp2,'Threshold',0.8);
annotations = string(labels) + ": " + string(scores);
%loop through each detected object
no_rows = height(bboxes);
for i=1:no_rows
    label = labels(i);
    bbox = bboxes(i,:);
    score = scores(i);
    
    if label=='Cancer'
        color = 'red';
    elseif label == 'Normal'
        color = 'green';
        
        I = insertObjectAnnotation(newp2,'rectangle',bboxes,cellstr(annotations),'Color',color);
    end
end
figure; imshow(I);

[Sachin]

Hi

Can you please tell what kind of error you are getting?

[Fahmi Akmal Dzulkifli]

The error is the colour of the bounding boxes are same for all cells. Like in the above example, all the cells have green bounding boxes colour, includes the cancer cells. When I checked at the 'labels', it has two classes, but when to assign the colour, it only has one class only.