Image database encoding using HYBRID video encoder

How to compress a set of images in an efficient way is one of the most popular research topics. Currently, one of the most preferred standards for compressing still images is JPEG2000 which mainly utilizes Discrete Wavelet Transformation. The video compression standard AVC/H.264, however, can also be applied to compress a set of still images. It has been proved that H.264 Intra Mode has similar or even better compression performance than JPEG2000 in terms of PSNR. This paper proposes an innovative approach that mainly depends Inter Prediction Mode of H.264, to compress a set of still images. Considering different compression performances for different image sequences, this paper also proposes an algorithm to determine the optimal sequence of images that helps Inter Prediction Mode to gain better performance. Experimental results reveal that H.264 Inter Prediction Mode outperforms JPEG2000 in terms of PSNR. The gain of H.264 can be further enhanced when the sequence of images is rearranged according to the images sequencing algorithm proposed in this paper.