is an ImageSeries that contains an additional link to an ImageSeries that is masked. This type appears to be designed to store an alpha mask applied to a visual stimulus, and the mask changes over time. Despite being only a transparency mask, the doc says that it stores values as RGBA, not just A. It's not clear what those other RGB values would be. It's also not clear how this type would be used. Often, a user wants to store the stimulus that the subject saw, which if masked, would be the result of mask + original image. Perhaps if a user showed the same image multiple times at different levels of transparency or different locations of masking, this neurodata type would be useful, but that is a very specific experimental paradigm. ImageMaskSeries is not documented on PyNWB or MatNWB, and is not used on DANDI (probably in large part because it is not documented). Since NWB 2.0, the NWB team has not heard of a need for ImageMaskSeries.
Overall, I think ImageMaskSeries does not have a strong use case in modern usage, so to remove cruft and maintenance costs, I propose we deprecate ImageMaskSeries.
As discussed at the NWB meeting today. The
ImageMaskSeriesneurodata typehttps://github.com/NeurodataWithoutBorders/nwb-schema/blob/dev/core/nwb.image.yaml#L119-L129
is an
ImageSeriesthat contains an additional link to anImageSeriesthat is masked. This type appears to be designed to store an alpha mask applied to a visual stimulus, and the mask changes over time. Despite being only a transparency mask, the doc says that it stores values as RGBA, not just A. It's not clear what those other RGB values would be. It's also not clear how this type would be used. Often, a user wants to store the stimulus that the subject saw, which if masked, would be the result of mask + original image. Perhaps if a user showed the same image multiple times at different levels of transparency or different locations of masking, this neurodata type would be useful, but that is a very specific experimental paradigm.ImageMaskSeriesis not documented on PyNWB or MatNWB, and is not used on DANDI (probably in large part because it is not documented). Since NWB 2.0, the NWB team has not heard of a need forImageMaskSeries.Overall, I think
ImageMaskSeriesdoes not have a strong use case in modern usage, so to remove cruft and maintenance costs, I propose we deprecateImageMaskSeries.