We propose context-based adaptive in-loop filters not only to reduce complexity but also to achieve coding efficiency by combining deblocking and in-loop filters. The proposed adaptive filters are designed based on contexts such as slice types, boundary strength values, pixel positions, and boundary directions in order to improve coding gain. The proposed method is also embedded with a deblocking filter to reduce complexity and to maintain subjective quality. Experimental results show coding gains of approximately 6.3% and 0.5% compared with the deblocking filter of H.264/AVC and the conventional block-based adaptive loop filter, respectively. Then the proposed method reduces complexity by about 20.8%.