it takes a lot of work to simplify footage that much and still have it be recognizable.
If you really want to use footage, some tools that might help you are:
1. Filter > Pixilate > Mosaic
Mosaic pixilates your image into squares according to the averages of the pixels in each square. If you use it on real footage, it will make it look blurry and unrecognizable–you have many more steps ahead.
AE also has most of these filters if you plan to do it there.
2. Image > Adjustments > Posterize
Posterize will reduce the number of colors in your image. You don’t have a lot of control over which colors are selected, though, so be prepared to replace them.
3. Filter Gallery… Cutout
I like cutout more than posterize, since it reduces the number of colors to a pallete more like that of the original… no idea how this would work over many frames of moving footage. You’d have to do it before the mosaic, or do the mosaic again when you’re done.
4. Layer > Layer Style > Drop Shadow
Once you have a small number of colors, you’ll need each color on its own layer. if each layer has a drop shadow it will look like they’re sitting on top of each other. Real 3-D layers are annoyingly, impossibly hard to do in Photoshop.
HOWEVER,
All told I think it would be easier to do a crude animation and then use effects to make it look blocky–to me that looks like what happened here, no matter what anyone says. (who moves their arms like that when they drum??)
–fill a layer with the color you want
–add a “hide all” layer mask
–use a white brush to draw a body part on the layer mask
–run Mosaic on the layer mask (Filter > Pixilate > Mosaic)
–run Threshold on the layer mask. (Image > Adjustments > Threshold)
–add a drop shadow to the layer.