If free oxygen exists, as dissolved in sea water, then it can in principle be extracted easily --- the gills of fish do so all the time. If oxygen is not free but bound in an oxide, then decomposition of most oxides is endoergic, requiring free-energy input. A catylist usually does not change the net free-energy difference between reactants and products, but it may reduce the free-energy cost of traversing a high-free-energy transition state by providing a reaction mechanism through a lower-free-energy transition state. But some oxides decompose at moderate temperatures to yield free oxygen without a catalyst, for example barium peroxide BaO2 and mercuric oxide HgO. At somewhat lower temperatures barium oxide BaO will pick up oxygen from the air to yield the peroxide BaO2, and the cycle can then be repeated. This was once a common method of producing oxygen, before liquid oxygen was readily available.
As Jack stated, you will have to provide free energy to drive the liberation process. You can produce oxygen by electrolysis of water. This avoids the use of potentially dangerous compounds, but the energy efficiency will be worse.