Hyunki Baik

A center-biased hybrid search method using plus search pattern for block motion estimation

This paper proposes a new method for block motion estimation based on center-biased hybrid search (CBHS) using plus search that employ a hybrid of a compact plus shaped search and a diamond search. It has only 5 search points for the best case and, on average, 6.6 search points. CBHS is about 34.3 times faster than the full-search algorithm. Hence, CBHS is consistently faster than the other suboptimal block-matching techniques. This paper compares the popular suboptimal block-matching technique with CBHS in which both the processing speed and the accuracy of motion compensation are tested over widely used H.263 test video sequences.

New fast successive elimination algorithm

Presents a very fast exhaustive search algorithm for motion estimation. This algorithm is based on successive elimination algorithm (SEA) and fast algorithms for the estimation of motion vectors [2]. A new motion estimation method called new fast successive elimination algorithm (NFSEA) finds the same motion vectors as exhaustive search algorithm with far fewer computational load. NFSEA effectively eliminates the search points within the search window and thus decreasing the number of matching evaluations that require very intensive computations. So, NFSEA dramatically reduces the number of operations for finding motion vectors. Experimental results show that NFSEA is a very efficient algorithm for the estimation of motion vectors.

Nobel successive elimination algorithms for the estimation of motion vectors

We present fast algorithms to reduce the computation of block matching algorithms for motion estimation in video coding. Nobel successive elimination algorithms for the estimation of motion vectors are based on the successive elimination algorithm for motion estimation and fast algorithms for the estimation of motion vectors. Nobel successive elimination algorithms effectively eliminate the search points within the search window and thus decrease the number of matching evaluation that requires very intensive computations. The efficiency of each proposed algorithm was verified by experimental results.

An efficient multicast routing protocol for mobile hosts

We present an alternative design, Eff-MM (efficient mobile multicast). For efficiently supporting multicast for mobile hosts on the Internet. Providing multicast services to hosts has become popular and many multicast protocols have been proposed for mobile hosts. In the delivery tree problems occur whenever a member moves. Such problems are non-optimal delivery route and overheads caused by the frequent reconstruction of a multicast routing tree. To solve problems like this, the paper used the tunneling range and search method to routing paths around migration places with a simple message. The result provides the shortest routes for delivery of multicast datagrams to mobile hosts and reduces frequent reconstruction of the multicast tree reconfiguration.

Advanced Multilevel Successive Elimination Algorithms for Motion Estimation in Video Coding

In this paper, we present advanced algorithms to reduce the computation cost of block matching algorithms for motion estimation in video coding. Advanced Multilevel Successive Elimination Algorithms are based on Multilevel Successive Elimination Algorithm for Block Matching Motion Estimation [1] and Successive Elimination Algorithm [2]. Advanced Multilevel Successive Elimination Algorithms consist of three algorithms. The second algorithm is useful not only for Multilevel Successive Elimination Algorithm but also for all kinds of block matching algorithms. The efficiency of the proposed algorithms was verified by experimental results.

FRD-FPRED: a novel re-decode based error compensation method using fast re-decoding and fast prediction algorithm

The study aims to find an efficient compensation method of transmission errors in low bit-rate mobile communications using the H.263 video codec. We suggest a novel error compensation method in the encoder (or transmitter), that can remove visual quality degradation due to spatio-temporal error propagation by utilizing a feedback channel. Our proposed method can reduce the computational complexity of the re-decoding method and FRD (fast re-decoding) method by using the characteristic for non-error-propagated macroblocks. Experimental results show that the proposed error compensation method can give higher coding efficiency compared with the other two non-re-decoding based methods: the reference picture selection (RPS) method and the error tracking (ET) method. Finally, our proposed method requires only 30% of the operations compared to the fast re-decoding (FRD) procedure.