Media Server image classification uses Convolutional Neural Network (CNN) classifiers. A CNN classifier usually produces more accurate results than other types of classifier, but can require a significant amount of time to train.
The more time you allow Media Server to train the classifier, the greater the accuracy. Before you train a CNNclassifier, you can choose how many training iterations to run. The time required to train the classifier is proportional to the number of training iterations and the number of training images. Increasing the number of iterations always improves the training and results in better accuracy, but each additional iteration that you add has a smaller effect.
For classifiers that have four or five dissimilar classes with around 100 training images per class, approximately 500 iterations produces reasonable results. This number of iterations with this number of training images requires approximately three hours to complete on a CPU or five minutes to complete with a GPU. Micro Focus recommends a larger number of iterations for classifiers that contain many similar classes. For extremely complex classifiers that have hundreds of classes, you might run 200,000 training iterations. Be aware that running this number of training iterations with large numbers of training images on a CPU is likely to take weeks.
To find the optimum number of iterations, Micro Focus recommends that you start with a small number of iterations. Double the number of iterations each time you train, until classification accuracy is acceptable.
When you run classification, the classifier outputs a confidence score for each class. These scores can be compared across classifiers, and you can set a threshold to discard results below a specified confidence level.
The performance of classification is generally better if:
- the classifier contains only a few classes (but it must contain at least two classes).
- the classes are dissimilar. For example, when training a 'field' class and a 'beach' class, the presence of clouds in the sky in both sets of training images might cause confusion between the classes.
- the classes are trained with many images. Usually around 100 images are sufficient to train a class. If the images in a class are very similar, fewer images might be sufficient.
- the training images are representative of the variation typically found within the class. For example, to train a "dog" class, use images of dogs of different sizes, breeds, colors, and from different viewpoints.
- the training images contain little background or clutter around the object in the image.
-
the longest dimension (width or height) of the training image is at least 500 pixels - smaller images might result in reduced accuracy.
TIP:
High-resolution images where the object covers a small proportion of the image make poor training images. If you have a large image showing the object and it can be cropped such that its longest dimension still exceeds 500 pixels, Micro Focusrecommends cropping the image. If you crop an image, leave a gap around the object of at least 16 pixels.
| Send documentation feedback to Micro Focus |
