Update ImageView to include layout and render filter operations

[dhrogers]

Adding a table input and designating column values for things like render filter (to be used to highlight), and a layout operation.

Comparative visualization can be done with addition of a single column, and a render attribute.

Proposed design of a table-based data flow that can be used by filters to support highlighting and comparative visualization by providing columns as input attributes to rendering and visualization operations.

Proposed extension of view objects to a common base class that can show any kind of visualization object. In this example, both image and plot views are laid out using the same input data, to show the relationship between the objects (and underlying data) in the views.

[dhrogers]

@JonasLukasczyk, @dbanesh - considering selection and highlighting as columns in an input table, we can support flexible comparative visualization through the filter graph. See diagram.

Could be used in the 'ML Model' workflow, as a 'classification' column would be added by Divya's MLModelFilter. That column can then be used by an ImageView to show how objects (in this case images) are being classified by the model. A later ImageView could remove items, but preserving the layout and using selection (as in the example above) shows the user what the classification operation is doing to the input data.

Within this framework, it is also easy to layout the images according to some non-grid layout. For example, the x and y axes of the layout could be thought of as a graph, and the locations of the images could show (for example) classification groups, etc. using columns in the input table as values.