Matlab Problem with dot indexing

So I want to train a Deep Learning network for Lane Detection from this Repo (https://github.com/SeokjuLee/VPGNet) and there is a matlab file which I need to run for some annotations.

Unfortunately I am getting an error:

Error in vpg_annot_v1 (line 42)

numFrames = size(gLabelData.frames,2);

Dot indexing is not supported for variables of this type.

What do I have to do to get this script working?

The full Code is:

%%2017.12.18. Seokju Lee & Junsik Kim
% Matlab code to convert caltech annot. into VPGNet annot.
% Usage: Download Caltech Lanes dataset and change path & dir parameter 
% <Directory structure>
% |__ caltech-lanes-dataset (download dataset[1])
%     |__ caltech-lane-detection/matlab (copy 'caltech-lane-detection/matlab' [2])
%     |__ cordova1
%     |__ cordova2
%     |__ washington1
%     |__ washington2
%     |__ vpg_annot_v1.m (current code)
%     |__ output.txt (output list file)
%
%   [1] http://www.mohamedaly.info/datasets/caltech-lanes
%   [2] https://github.com/SeokjuLee/caltech-lane-detection

clear all; 
close all; clc;

%%startup (change category)
category = 'cordova1';
% category = 'cordova2';
% category = 'washington1';
% category = 'washington2';
addpath(genpath('./caltech-lane-detection/matlab'))

path = '/home/orkun/VPGNet/caltech-lanes-detection';                      % change path
file =  sprintf('/%s/labels.ccvl', category);
gLabelData = ccvLabel('read', [path file]);

%%spline & grid mask parameters 
h = 0.02;           % interval for splines
height = 480;       % reference image height
width = 640;        % reference image width
gg = 8;             % grid size for binary mask
grid_x = 1:gg:481;
grid_y = 1:gg:641;
thickness = 2;

%%make annotation file
fileID = fopen(sprintf('./%s.txt', category),'w');
numFrames = size(gLabelData.frames,2);
gLabelSubtypes = {'bw', 'sw', 'dy', 'by', 'sy'};

for i = 1:numFrames
    disp(sprintf('frame: %03d',i))
    numLanes = size(gLabelData.frames(i).labels, 2);
    segs = [];
    splines = {numLanes};
    sptypes = {numLanes};
    for j = 1:numLanes
        splines{j} = ccvEvalBezSpline(gLabelData.frames(i).labels(j).points, h);     % convert 4 point to spline
        sptypes{j} = gLabelData.frames(i).labels(j).subtype;
        splines_x1 = splines{j};
        splines_x2 = splines{j};
        splines_x1(:,1) = splines_x1(:,1) - thickness;
        splines_x2(:,1) = splines_x2(:,1) + thickness;
        splines{j} = [splines{j}; splines_x1; splines_x2];

          for k = 1:size(splines{j},1)    % make spline points into bounding box
              grid_pos_x = floor((splines{j}(k,1)-1)/gg);
              grid_pos_y = floor((splines{j}(k,2)-1)/gg);
              xmin = grid_pos_x * gg + 1;
              xmax = grid_pos_x * gg + gg;
              ymin = grid_pos_y * gg + 1;
              ymax = grid_pos_y * gg + gg;
              grid_width = xmax - xmin;
              grid_height = ymax - ymin;
              inst_id = j;
              lane_id = find(ismember(gLabelSubtypes, sptypes{j}));
              segs = [segs; xmin, ymin, xmax, ymax, inst_id, lane_id];
          end
          segs = unique(segs, 'rows');
      end    
      numLaneSegs = size(segs, 1) - numLanes;
      fprintf(fileID,'/%s/f%05d.png  %d',category, i-1, size(segs,1));

%     img_path = sprintf('%s\\f%05d.png',category,i-1);
%     img = imread(img_path);
%     figure(99);
%     imshow(img);
%     hold on;

      for j = 1:size(segs,1)    
          xmin = segs(j,1);
          ymin = segs(j,2);
          xmax = segs(j,3);
          ymax = segs(j,4);
          inst_id = segs(j,5);
          lane_id = segs(j,6);

          fprintf(fileID, ' ');

          fprintf( fileID, ' %d', xmin );
          fprintf( fileID, ' %d', ymin );
          fprintf( fileID, ' %d', xmax );
          fprintf( fileID, ' %d', ymax );
  %         rectangle('Position', [xmin ymin xmax-xmin ymax-ymin])

          fprintf( fileID, ' %d', lane_id ); % depth data -> lane_id        
      end

%     waitforbuttonpress;

      fprintf(fileID,'\n');    

      %%Visualize
  %     img_path = sprintf('%s\\f%05d.png',category,i-1);
  %     img = imread(img_path);
  %     mask_img = zeros(size(img,1), size(img,2));
  %     
  %     fig = figure(1);
  %     set(fig, 'position', [0, 0, 2000, 1000]);
  %     subplot(1,2,1);
  %     imshow(img);
  %     hold on;
  %     for j = 1:numLanes
  %         plot(splines{j}(:,1), splines{j}(:,2),'b.','markersize',6);
  %     end
  %     hold off
  %     
  %     subplot(1,2,2);
  %     imshow(img);
  %     hold on;
  %     for j = 1:size(segs,1)
  %         rectangle('Position',[segs(j,1),segs(j,2),gg,gg]);
  %     end
  %     hold off;
  %     
  %     waitforbuttonpress;

      %%Make lane GT
  %     global_im=zeros(480,640);
  %     locc=segs(:,1:4);
  %     for ii=1:size(locc,1)
  %         im=zeros(480,640);
  %         temp=[locc(ii,1) locc(ii,2) locc(ii,3) locc(ii,2) locc(ii,3) locc(ii,4) locc(ii,1) locc(ii,4)]; 
  %         res=im2bw(insertShape(im,'FilledPolygon', temp));
  %         global_im=global_im|res;
  %     end
  %     ver = imresize(global_im,[60,80]);
  %     imwrite(ver,['./gt_caltech/' category '/' gLabelData.frames(i).frame]);

end

fclose('all');

Answer 1

[Stephen23]

This happens because gLabelData is not a structure, which you code expects it to be.

So why is gLabelData not a structure?

Unfortunately ccvLabel is not very well written code. It is certainly not robust code. For example, just within the small part related to file importing, it:

1. load-s directly into the workspace (ugh, this is a very bad idea)...
2. which forces the author to use slow exist instead of efficient isfield.
3. If the load operation throws an error, then it simply gets ignored and the function returns nothing:

try
  load(fname, '-mat');
catch
  return;
end;

For example, if the file cannot be found (e.g. spelling mistake, wrong directory, etc), then this code will not give any error or warning about this and will just return as if nothing had happened. This is a terrible approach to writing code. Do NOT learn from this badly thought out code (even if it was written by someone at CalTech... academics can write bad code).

In any case, I suspect that that is what is happening in your case: the filename/filepath you are using does not match any file on your file system.

Solution: use exist to check the filename that you are using, before supplying it to this to ccvLabel.