Platform for learning based recognition research

A computing environment, comprising using one or more computers to perform process actions to:

• gather one or more data samples from each of a plurality of users in the computing environment via a user interface into a plurality of training data sets used for creating a recognition model;
• store the training data sets in the cluster of computers;
• receive a selection of training parameters via the user interface for use in constructing the recognition model by:
• selecting one or more feature extraction algorithms from a display of a plurality of feature extraction algorithms, each feature extraction algorithm configured to extract one or more features from data samples,
• selecting one or more classifier algorithms from a display of a plurality of classifier algorithms, each classifier algorithm configured to associate extracted features to a template set, and
• selecting one or more of the training data sets from a display of the plurality of training data sets;
• create the recognition model based on the selection of training parameters;
• and
• evaluate the recognition model to determine an accuracy of the recognition model.

A method for researching and developing a recognition model in a computing environment, comprising:

• providing a user interface for displaying a plurality of subsets of a training data set, each subset having one or more data samples, the subsets having been previously defined by one or more users in the computing environment;
• selecting one or more of the subsets via the user interface;
• selecting one or more feature extraction algorithms that were previously uploaded by the users from a plurality of feature extraction algorithms displayed via the user interface, each feature extraction algorithm being configured to extract one or more features from each data sample of the selected subsets;
• selecting one or more classifier algorithms that were previously uploaded by the users from a plurality of classifier algorithms displayed via the user interface, the classifier algorithms being configured to associate the features to a template set;
• creating the recognition model from the selected feature extraction algorithms, the selected classifier algorithms, and the selected subsets;
• and
• evaluating the recognition model to determine an accuracy of the recognition model.