Avi Nehemiah is the product marketing manager for computer vision applications.
Avi's pick of the week is
imageSet viewer by
Brett Shoelson.
My pick of the week is the
imageSet Viewer user interface created by my friend Brett Shoelson. The biggest challenges I face when working with large sets of images(that are normally stored in several different directories) is it was difficult to bring the data into MATLAB, and interactively visualize the images.
The imageSet functionality in the
Computer Vision System Toolbox helps me bring the data into MATLAB. I can bring all the images in a folder into MATLAB, and maintain the hierarchical relationship between folders using the imageSet. The image data I am using is from
Caltech 101, collected by Fei-Fei Li, Marco Andreetto, and Marc 'Aurelio Ranzato. imageSet functionality is new in R2014b, you can learn more about it in the
documentation .
imageData = imageSet('Data','recursive');
Now let's take a look at some of properties of the imageData variable I just created.
numImages = [imageData.Count]'
folderDescriptions = {imageData.Description}'
numImages =
17
8
2
2092
2292
472
790
770
118
folderDescriptions =
'InteractiveTest'
'InteractiveTest_NoiseCompare'
'InteractiveTest_TestClassifier'
'Motorbikes'
'airplanes'
'car_side'
'Motorbikes'
'airplanes'
'car_side'
This is where the imageSet Viewer comes in to help me interactively visualize this data
imageSetViewer(imageData)
Other options to navigate through your images using the user interface are
- Click on the tabs at the top to view different folders or categories
- Click on an image to view in a separate window
- Right click on an image to save to workspace
- Automatically creates imageSet if you pass in the directory name as an argument
- Click on the name of an image to copy to clipboard
I'd recommend trying out the imageSet viewer anytime you are working with sets of images.
Comments
To leave a comment, please click here to sign in to your MathWorks Account or create a new one.