isolating bounding boxes from a video using detector

Hi mathworks team,

I am trying to isolate bounding boxes values from frames of video using trained detector but it gives this error ''Index exceeds the number of array elements. Index must not exceed 1.''

here is code

% import the video file

obj = VideoReader('test2video.mp4');

vid = read(obj);

% read the total number of frames

frames = obj.NumberOfFrames;

% file format of the frames to be saved in

ST ='.jpg';

a = cell(1, frames+1);

% reading and writing the frames

for x = 1 : frames

% % converting integer to string

Sx = num2str(x);

%

% % concatenating 2 strings

Strc = strcat(Sx, ST);

Vid = vid(:, :, :, x);

cd frames;

% %

% % % exporting the frames

imwrite(Vid, Strc);

cd ..

for i = 1:x

m = imread(['C:\Users\shirs\Documents\SURF\SSDVOBJECT\frames\' Sx(i) '.jpg']);

[bboxes, scores] = detect(detector, m);

a{i} = bboxes;

figure

imshow(m)

end

end

[Nikhilesh]

In this particular part what i belive might be the issue.

imwrite(Vid, Strc);
     cd ..

[Shirshak]

No .. this line is for saving it in one folder..then from there i try to run a loop and isolate the bounding boxes.

any answers ?

[Shirshak]

hi team, did you get any answers on this

Answer 1

[Ajay Gajulapally]

Hi Shirshak,

As per my understanding, you want to isolate the bounding boxes of detected objects in a video frame by frame. Kindly check the way you use your second for loop. The modified code can go as:

1. Importing the video file and reading the frames

% import the video file
obj = VideoReader('________'); % use the video file here
vid = read(obj);

 % read the total number of frames
frames = obj.NumFrames;

2. Writing the frames to a required path.

% file format of the frames to be saved in
ST ='.jpg';
% reading and writing the frames 
for x = 1 : frames
%     % converting integer to string
   Sx = num2str(x);
%   
%      % concatenating 2 strings
      Strc = strcat(Sx, ST);
     Vid = vid(:, :, :, x);
     cd frames;
     
% %     % exporting the frames
     imwrite(Vid, Strc);
     cd ..
end

3. Load your trained detector and save the bounding boxes to a variable.

 detector = yolov3ObjectDetector("tiny-yolov3-coco"); % load your trained detector here
     a = cell(1,frames+1);
     for i = 1:frames
         z = "PATH_NAME\frames\"+string(i)+".jpg";
          m = imread(z);
          [bboxes, scores] = detect(detector, m);
           a{1,i} = bboxes;
            annotatedImage = insertShape(m, 'Rectangle', a{i});
            figure
            imshow(annotatedImage);
     end

Hope this helps!