In an uploaded data table, each object/ROI must be on one row but to link this row to the pixel coordinates of a mouse click on a corresponding image, the ROIs coordinates should be represented in the table. ROIs are commonly represented by centre of mass or bounding box. 1- Which of these two representations should we support or should we support both? 2- Which form should this representation take in the data table?
I would suggest to go for the simplest case which is centre of mass coordinates (x, y, z, channel, time) either in one column or across 5 columns.