predictFace.m

function selectedBbox = predictFace(img_rz, matfile) 
%#codegen

% This function detects the traffic signs in the image using Detection Network
% (modified version of Yolo) and recognizes(classifies) using Recognition Network
%
% Inputs :
%
% im            : Input test image
%
% Outputs :
%
% selectedBbox  : Detected bounding boxes 
% idx           : Corresponding classes

% Copyright 2017-2018 The MathWorks, Inc.

coder.gpu.kernelfun;

% Converting into BGR format
img_rz = img_rz(:,:,3:-1:1);
img_rz = im2single(img_rz);

%% TSD
persistent detectionnet;
if isempty(detectionnet)   
    detectionnet = coder.loadDeepLearningNetwork(matfile,'Detection');
end

predictions = detectionnet.predict(img_rz);%, 'executionenvironment', 'cpu');

%% Convert predictions to bounding box attributes
classes = 1;
num = 2;
side = 11;
thresh = 0.2;
[h,w,~] = size(img_rz);

boxes = single(zeros(0,4));    
probs = single(zeros(0,1));    
for i = 0:(side*side)-1
    for n = 0:num-1
        p_index = side*side*classes + i*num + n + 1;
        scale = predictions(p_index);       
        prob = zeros(1,classes+1);
        for j = 0:classes
            class_index = i*classes + 1;
            tempProb = scale*predictions(class_index+j);
            if tempProb > thresh
                
                row = floor(i / side);
                col = mod(i,side);
                
                box_index = side*side*(classes + num) + (i*num + n)*4 + 1;
                bxX = (predictions(box_index + 0) + col) / side;
                bxY = (predictions(box_index + 1) + row) / side;
                
                bxW = (predictions(box_index + 2)^2);
                bxH = (predictions(box_index + 3)^2);
                
                prob(j+1) = tempProb;
                probs = [probs;tempProb];
                                
                boxX = (bxX-bxW/2)*w+1;
                boxY = (bxY-bxH/2)*h+1;
                boxW = bxW*w;
                boxH = bxH*h;
                boxes = [boxes; boxX,boxY,boxW,boxH];
            end
        end
    end
end

%% Run Non-Maximal Suppression on the detected bounding boxess
coder.varsize('selectedBbox',[98, 4],[1 0]);
[selectedBbox,~] = selectStrongestBbox(round(boxes),probs);

% Copyright 2020 The MathWorks, Inc.