With dimension reduction, It's possible to visualize image data in 2D by using the PlotMarkers option:
ListPlot[List /@ xy,
PlotMarkers -> (Image[#, ImageSize -> 100] & /@ tree),
ImageSize -> 2000]
(which comes from here: https://www.wolfram.com/language/11/improved-machine-learning/visualize-a-dataset-using-feature-extraction.html?product=mathematica)
However, if one wishes to have 3D dimension reduction with images as points, you can't really do that.
I have seen a post that says one can replace the points in ListPointPlot3D with shapes like so:
lpdata = Table[(4 Pi - t) {Cos[t + Pi/2], Sin[t + Pi/2], 0} + {0, 0, t}, {t, 0, 4 Pi, .1}];
lpp1 = ListPointPlot3D[lpdata,
Filling -> Bottom, ColorFunction -> "Rainbow", BoxRatios -> 1,
FillingStyle -> Directive[LightGreen, Thick, Opacity[.5]], ImageSize -> 400];
lpp2 = lpp1 /. Point[x__] :> (Sequence@{EdgeForm[], Cone[#, .3]} &@
({x} /. {{a_, b_, c_}} :> {{a, b, c}, {a, b, .5 + c}}));
Row[{lpp1, lpp2}, Spacer[5]]
From: Point style in ListPointPlot3D
This and this post: Use a custom3D graphic for points in ListPointPlot3D
Are somewhat relevant, but I wanted to ask if it's at all possible to replace points with classic 'sprites' that are 2D images in 3D space that always face the viewer, or even just small static 2D planes in 3D space with the image projected onto it would be great.
Is any of this possible?
Inset[]:Graphics3D[Table[Inset[Graphics[{Texture[ExampleData[{"TestImage", "Lena"}]], Polygon[{{0, 0}, {1, 0}, {1, 1}, {0, 1}}, VertexTextureCoordinates -> {{0, 0}, {1, 0}, {1, 1}, {0, 1}}]}, ImageSize -> Tiny], ConstantArray[j, 3]], {j, -4, 4, 2}]]– J. M.'s missing motivation Dec 30 '16 at 09:24Normal[lpp1] /. Point[plist__] :> Inset[sprite, #] & /@ plist– J. M.'s missing motivation Dec 30 '16 at 10:23