In this paper a 3-module automatic watermarking system is proposed, which embeds visually recognizable patterns to video objects (VOs). During the first module efficient unsupervised video object extraction is performed. In the second module the DWT is applied to the extracted VOs to produce ten subbands and three pairs of subbands are formed. Next QSWTs are estimated for the pair of subbands with the highest energy content. Finally in the third module visually recognizable patterns are embedded to the coefficients of the highest energy QSWTs and the IDWT is applied to provide the watermarked VO. The performance of the proposed system has been tested under various signal distortions producing very promising results.