If you reach an error "Undefined function 'testSift'
then you need to find all occurances of this method and uncomment the mlpfwd
method instead:
[mPrediction,~]=testSift(trainingSet,predictImage); // replace this ...
mPrediction = mlpfwd(mNet,mFrame); // ... with this
You'll then need to get mlpfwd
from here
This is the result of an incomplete refactor.
=========================================================================== 24-787 Project: American Sign Language (ASL) Recognition System
The purpose of this document is to provide useful instructions and details pertaining to the MATLAB code developed for this project.
%%%%% ACQUIRING TRAINING DATA %%%%% To get training data for the neural network, follow these steps:
% Load the GUI
>>aslgui
% Set the directory in which to store the training samples
% (default is '/training data' subfolder of current directory)
% Set the desired filename prefix. For now, I propose to
% use the initials of the subject. Currently, the filename convention
% is fixed to PRE###.mat where PRE is the prefix and ### increments for
% each sample.
% Turn the camera on by pressing the START CAMERA button.
% The camera should begin streaming in the VIDEO STREAM panel.
% Make sure that the desired class (letter sign) is selected in the
% PALETTE panel. The letter should appear below along with a
% corresponding value (1-26 for A-Z).
% When you are ready, press the CAPTURE FRAME button to save a single
% frame of data. The frame is saved as a variable DATA in the selected
% directory with the selected filename. DATA is a 307201x1 vector,
% where the first
To view a complete listing of video input functions and properties, use >>imaqhelp videoinput
Example:
% Construct a video input object
obj = videoinput('winvideo');
% Select the source to use for acquisition.
set(obj, 'SelectedSourceName', 'input1')
% View the properties for the selected video source object.
src_obj = getselectedsource(obj);
get(src_obj)
% Preview a stream of image frames.
preview(obj);
% Acquire and display a single image frame.
frame = getsnapshot(obj);
image(frame);
% Remove video input object from memory.
delete(obj);
videoaq.m: acquire video via webcam and display in a GUI
Notes:
File list:
- aslgui.m: The main file containing a GUI for acquiring training data.
- signPalette.m: The file that contains the list of available signs (nested in aslgui.m).