Online ISSN
1751-9683
Print ISSN
1751-9675
IET Signal Processing
Volume 1, Issue 3, September 2007
Volumes & issues:
Volume 1, Issue 3
September 2007
-
- Author(s): H.-M. Wang ; C.-H. Tseng ; J.-F. Yang
- Source: IET Signal Processing, Volume 1, Issue 3, p. 121 –127
- DOI: 10.1049/iet-spr:20065007
- Type: Article
- + Show details - Hide details
-
p.
121
–127
(7)
In H.264 advanced video coding standard, the intra prediction made by referring to the surrounding coded pixels plays an important role in achieving the compression efficiency. To find the best prediction mode, the H.264 reference software JM 6.1d suggests two criteria, the sum of absolute differences (SAD) and the sum of absolute transformed differences (SATD). The SATD criterion with extra Hadamard transform achieves better coding performance, but needs more computation than the SAD criterion. A fast algorithm is developed to reduce the computation of the SATD, by using the property of linear transform and the fixed-spatial relationship of predicted pixels in each intra mode. With the fast SATD computation, about 50% computation of the original SATD algorithm can be reduced. To further reduce the SATD computation, a two-stage simplified SATD method is proposed, which first uses the SAD criterion to eliminate some unwanted modes and then performs the fast SATD computation to select the best mode. Experimental results show that the proposed two-stage simplified SATD method can save about 70% of computation costs with negligible PSNR loss for H.264 intra prediction. - Author(s): K.J. Jones and R. Coster
- Source: IET Signal Processing, Volume 1, Issue 3, p. 128 –138
- DOI: 10.1049/iet-spr:20060329
- Type: Article
- + Show details - Hide details
-
p.
128
–138
(11)
A highly parallel hardware solution to the regularised fast Hartley transform (FHT), an algorithm developed only recently for the efficient computation of the discrete Hartley transform (DHT) and hence of the real-data fast Fourier transform (FFT), is discussed. A drawback of conventional FHT algorithms lies in the loss of regularity arising from the need for two types of ‘butterfly’ for catering for fixed-radix formulations. A generic radix-4 double butterfly was therefore developed for the regularised FHT which overcame the problem in an elegant fashion. An architecture for the parallel computation of the generic double butterfly and the resulting regularised FHT is now discussed in some detail; this exploits a single high-performance processing element (PE) which yields an attractive solution, particularly when implemented with field-programmable gate array (FPGA) technology, which is both area-efficient and scalable in terms of transform length. High performance is achieved by having the PE able to process the input/output data sets to the double butterfly in parallel, this in turn implying the need to be able to access simultaneously, and without conflict, both multiple data and multiple twiddle factors, or coefficients, from their respective memories. The resulting area-efficient and scalable single-PE architecture is shown to yield a generic solution to the real-data FFT which is capable of achieving the throughput of the most advanced commercially available solutions for just a fraction of the silicon area and, for new applications, at negligible re-design cost. - Author(s): P.-C. Lin ; J.-C. Wang ; J.-F. Wang ; L.-C. Wen
- Source: IET Signal Processing, Volume 1, Issue 3, p. 139 –149
- DOI: 10.1049/iet-spr:20060326
- Type: Article
- + Show details - Hide details
-
p.
139
–149
(11)
With the increasingly widespread use of personal portable devices, it is essential to devise an efficient method for spoken data retrieval for its resource-limited identity. This investigation proposes two efficient feature-based sentence-matching algorithms for speaker-dependent personal spoken sentence retrieval. Such a system can efficiently retrieve database sentences only partially matched to query sentence inputs. The query and database sentences are initially segmented into equal-sized matching units. A matching plane that comprises matching blocks is then created. For each matching block, a local similarity score is then determined from the feature distance. A whole-matching-plane-based accumulation scheme and a column-based row-based accumulation scheme are then designed to determine the global similarity score. The global similarity score of the matching plane reveals the similarity between the query and database sentences. The proposed algorithms are based on the feature-level comparison and do not require acoustical and language models. Experiments on news titles and personal schedules were conducted. The experimental results show that the proposed algorithms can efficiently work on both PC and HP iPAQ H5550 PDA. - Author(s): W.-G. Chen and Y. Ling
- Source: IET Signal Processing, Volume 1, Issue 3, p. 150 –155
- DOI: 10.1049/iet-spr:20070003
- Type: Article
- + Show details - Hide details
-
p.
150
–155
(6)
In a practical video encoder, a video sequence obtained from a camera inevitably conveys noise. The noise term degrades not only image quality but also coding efficiency. Based on the statistical analysis of noise signal, a two-stage noise variance adaptive successive elimination algorithm (SEA) for block motion estimation is presented. To reduce the additional computation cost required for video noise estimation, it is embedded into the process of block matching using Kalman filtering. Simulation results demonstrate that the performance of the proposed algorithm is close to that of the SEA, whereas the computational complexity has been significantly reduced. - Author(s): H.T. Hu and C. Yu
- Source: IET Signal Processing, Volume 1, Issue 3, p. 156 –163
- DOI: 10.1049/iet-spr:20070008
- Type: Article
- + Show details - Hide details
-
p.
156
–163
(8)
An adaptive scheme for noise spectral estimation is proposed to cope with the power spectral subtraction method for speech enhancement. The use of such a noise spectral estimate helps eliminate residual noise without increasing speech distortion. Improvements in the segmental signal-to-noise ratio and modified bark spectral distortion measures are evidently seen with 192 TIMIT sentences corrupted by four types of broadband noise taken from the Noisex-92 database. Moreover, incorporation of this noise adaptation scheme into the constrained parameter estimator is demonstrated. Mean opinion score listening tests confirm that the proposed subtractive scheme yields better results than that acquired by subtracting an averaged noise power spectrum. - Author(s): F. Yan ; L. Cheng ; H. Wang
- Source: IET Signal Processing, Volume 1, Issue 3, p. 164 –169
- DOI: 10.1049/iet-spr:20070028
- Type: Article
- + Show details - Hide details
-
p.
164
–169
(6)
A new family of dual-tree based dyadic complex (approximate analytic) wavelet tight frames that is an extension of higher density discrete wavelet transform (DWT) introduced by Selesnick (in A Higher-density DWT, IEEE Transaction on Signal Processing) is proposed. Because the proposed wavelet frames are good combinations of the higher-density DWT and the dual tree complex wavelet transform, the corresponding new wavelet transform can be potentially applied to the applications when properties of both transforms are required simultaneously. The design problem to obtain finite impulse response filters satisfying the numerous constraints imposed by this new wavelet transform is addressed. The proposed wavelet frames are compactly supported, nearly shift-invariant, having vanishing moments and intermediate scales. Several explicit examples are discussed to illustrate the construction and properties of the wavelet frames.
Computation reduction for intra 4×4 mode decision with SATD criterion in H.264/AVC
Area-efficient and scalable solution to real-data fast fourier transform via regularised fast Hartley transform
Design and portable device implementation of feature-based partial matching algorithms for personal spoken sentence retrieval
Noise variance adaptive successive elimination algorithm for block motion estimation: application for video surveillance
Adaptive noise spectral estimation for spectral subtraction speech enhancement
Higher density dual-tree discrete wavelet transform
Most viewed content for this Journal
Article
content/journals/iet-spr
Journal
5
Most cited content for this Journal
-
Parameter estimation algorithms for dynamical response signals based on the multi-innovation theory and the hierarchical principle
- Author(s): Ling Xu and Feng Ding
- Type: Article
-
Acoustic vector sensor: reviews and future perspectives
- Author(s): Jiuwen Cao ; Jun Liu ; Jianzhong Wang ; Xiaoping Lai
- Type: Article
-
Two-dimensional DOA estimation for L-shaped array with nested subarrays without pair matching
- Author(s): Yang-Yang Dong ; Chun-Xi Dong ; Ying-Tong Zhu ; Guo-Qing Zhao ; Song-Yang Liu
- Type: Article
-
Image super-resolution reconstruction using the high-order derivative interpolation associated with fractional filter functions
- Author(s): Deyun Wei
- Type: Article
-
Convolution and correlation theorems for the two-dimensional linear canonical transform and its applications
- Author(s): Qiang Feng and Bing-Zhao Li
- Type: Article