A significant portion of cave paintings are so called "hand stencils": People pressed one hand onto the wall and applied color around the hand.
Image: Cueva de las Manos, Perito Moreno, Argentina / Wikipedia
So from these paintings, we can infer the shape of the artist's hand. It had been known for a long time that among the hand stencils there were bigger and smaller hands. One older interpretation was that the bigger hands belonged to older, experienced (male) hunters, and the smaller hands belonged to young (male) "novices", maybe as some kind of intiation ritual.
But about ten or twenty years ago, scientists started to look closer into the hand shapes, and found interesting results. The proportions of a male hand and a female hand tend to be different. For example, with males the ring finger is typically significantly longer than the index finger. With females it's the other way around, or both fingers are of about equal length.
One of the first scientists to look into this was Dean Snow, an anthropologist from Pennsylvania State University. According to an German article on wissenschaft.de, the homepage of the German science magazine "bild der wissenschaft", he leafed through a photo book with pictures of stone-age cave art. There, he ran across a hand stencil that clearly didn't originate from a man, but from a woman. In that book alone, out of the six stencils shown, four were from female hands.
Snow decided to look closer into this, visited several caves with stone age art in Europe and got high-res photos from other caves. Additionally, he took measurements from the hands of contemporary people from the respective regions. He found that only about 10% of the stone age hand stencils had originated from adult males, and 15% from juveniles. The vast majority of 75% of the stencils clearly had originated from adult females. According to Snow's measurements, the sexual dimorphism of the human hand was even more pronounced 30.000 years ago than it is today. He had feared that because of the significant overlap today, it might be hard to assign the sex of the stencils' creators. But the stone age hands all fell into the extreme areas of the modern statistical distribution.
Snow published his findings in a paper in the journal "Antiquity" in 2006. Since, other scientists also found that at least the hand stencils weren't as much of a "male domain" as had been assumed before, for example Paul Pettitt (Durham University) et al., also in "Antiquity".
Obviously, this method won't work for other cave paintings like depictions of animals.