In this paper, we propose a new polygonal mesh geometry coding scheme based on a process of structuring by acquiring surrounding vertices of the polygonal mesh one layer at a time. The structuring process begins by se...
详细信息
In this paper, we propose a new polygonal mesh geometry coding scheme based on a process of structuring by acquiring surrounding vertices of the polygonal mesh one layer at a time. The structuring process begins by selecting the start vertex and proceeding by acquiring surrounding vertices of the polygonal mesh. As a result, we obtain a 2-D structured vertex table. Structured geometry data are generated according to the structured vertices and encoded by a multiresolution decomposition and space frequency quantization coding method. In our proposed scheme, the multiresolution decomposition uses the connectivity of the polygonal mesh. In addition, with a space frequency quantization coding scheme, we can reduce redundancies of decomposed coefficients at similar positions in different components of decomposition level. Experimental results show that the proposed scheme gives better coding performance at lower bit-rates than the usual schemes.
In this paper, we present an efficient region-based video coding scheme. The region partitioning is obtained using a minimum description length (MDL) formalism in order to optimize the compression. We compare this app...
详细信息
In this paper, we present an efficient region-based video coding scheme. The region partitioning is obtained using a minimum description length (MDL) formalism in order to optimize the compression. We compare this approach with results obtained by MPEG and H.261 coders and show the gain obtained by our approach.
We propose a novel approach for light field compression that incorporates disparity compensation into 4-D wavelet coding using disparity-compensated lifting. With this approach, we obtain the benefits of wavelet codin...
详细信息
We propose a novel approach for light field compression that incorporates disparity compensation into 4-D wavelet coding using disparity-compensated lifting. With this approach, we obtain the benefits of wavelet coding, including compression efficiency and scalability in all dimensions. Additionally, our proposed approach solves the irreversibility limitations of previous wavelet coding approaches. Experimental results show that the compression efficiency of the proposed technique outperforms current state-of-the-art wavelet coding techniques by a wide margin.
Self-supervised learning (SSL) has shown promise in learning representations of audio that are useful for automatic speech recognition (ASR). But, training SSL models like wav2vec 2.0 requires a two-stage pipeline. In...
详细信息
ISBN:
(纸本)9781728176055;9781728176062
Self-supervised learning (SSL) has shown promise in learning representations of audio that are useful for automatic speech recognition (ASR). But, training SSL models like wav2vec 2.0 requires a two-stage pipeline. In this paper we demonstrate a single-stage training of ASR models that can utilize both unlabeled and labeled data. During training, we alternately minimize two losses: an unsupervised masked Contrastive predictive coding (CPC) loss and the supervised audio-to-text alignment loss Connectionist Temporal Classification (CTC). We show that this joint training method directly optimizes performance for the downstream ASR task using unsupervised data while achieving similar word error rates to wav2vec 2.0 on the Librispeech 100-hours dataset. Finally, we postulate that solving the contrastive task is a regularization for the supervised CTC loss.
The experimental paradigm of embodiment illusions has greatly contributed to our knowledge on how the brain distinguishes self from other. Different types of illusions have provided empirical data for a theoretical fr...
详细信息
The experimental paradigm of embodiment illusions has greatly contributed to our knowledge on how the brain distinguishes self from other. Different types of illusions have provided empirical data for a theoretical framework describing multisensory integration of signals from both within and outside of the body, and predictive coding mechanisms to weigh those signals. However, most embodiment illusion studies lean heavily on visual sensory information as the main signal to establish the illusion. Few studies to date have explored non-visual embodiment illusions. These could potentially lead to a more thorough understanding of the underlying mechanisms of embodiment. In this study, we aim to approach an auditory embodiment illusion - more specifically a voice illusion - in an in-depth, structured way. We combined vibrotactile feedback on the throat with voice sounds, both articulated and non-articulated. Additionally, we measured interoceptive, proprioceptive and exteroceptive sensitivity. Results indicate that non-visual embodiment illusions are much more difficult to establish than visual-based ones and that proprioceptive and interoceptive sensitivity might influence the illusion strength. Absence of feedback might disrupt the illusion less than asynchronous feedback, which is in line with predictive coding expectations.
In this paper, we present a VQ-based two-codebook design which uses separate codebooks for predicted residuals and full pixel values. We show that our approach captures abrupt scene changes while exploiting inter-fram...
详细信息
In this paper, we present a VQ-based two-codebook design which uses separate codebooks for predicted residuals and full pixel values. We show that our approach captures abrupt scene changes while exploiting inter-frame dependencies. We use a simple universal code consisting of two codebooks, an intra-codebook containing codewords that are used as reproduction of image blocks together with a residual-codebook. Codebooks are selected to minimize distortion for the block being coded. If there is relatively little motion in a frame, most blocks use the residual-codebook. On the other hand, if a frame is very different from the previous one, most blocks will be coded using the intra-codebook. When compared to other VQ schemes mentioned above, the resulting quantizer not only follows scene changes closely with satisfactory fidelity but also is robust against mismatch between the training and test sequence. We compare the PSNR of three coding schemes, intra-coding-only, inter-coding only, and the proposed method.
Very sparse bitmaps are used in a wide variety of applications, ranging from adjacency matrices in representation of large sparse graphs, representation of sparse space occupancy to book-keeping in databases. A method...
详细信息
Very sparse bitmaps are used in a wide variety of applications, ranging from adjacency matrices in representation of large sparse graphs, representation of sparse space occupancy to book-keeping in databases. A method based on pruning of the binary space partition (BSP) tree in the minimal description length (MDL) principle for coding very sparse bitmaps was proposed. This new method for coding of sparse bitmaps meets seemingly competing objectives of good compression, the ability of conducting queries directly in the compression domain, and simple and fast decoding.
Lossless compression techniques-are essential in archival and communication of medical images. Compression of medical images has been historically reversible or "lossless" limiting compression ratios to betw...
详细信息
Lossless compression techniques-are essential in archival and communication of medical images. Compression of medical images has been historically reversible or "lossless" limiting compression ratios to between 2:1 and 4:1. A new and simple lossless method-interval number-in the field of image compression in medical science is presented. The method is tested on several kinds of popular medical images. The general compression ratio is around 2:1. The best compression ratio can be 7.8:1 for X-ray images. It is shown that the interval number method is a very efficient and powerful method for medical image compression.
Data compression is a trending field that is used in data storage and data transmission systems. Lossy compression means that data cannot be completely retrieved while in lossless compression the compressed data must ...
详细信息
ISBN:
(纸本)9781665482387
Data compression is a trending field that is used in data storage and data transmission systems. Lossy compression means that data cannot be completely retrieved while in lossless compression the compressed data must be reconstructed exactly. Lossless data compression is used in compressing binary files, telemetry data and high-fidelity medical and scientific images where details are crucial. There is no generic compression algorithm that gives best compression ratio on all data pattern. In this paper, we propose a hybrid lossless hardware architecture that compresses most of data patterns such as repeated data, Gaussian distribution data and images. A profiling-before-compressing and then choosing the right compression hardware is proposed. The proposed design is a highly parallelized architecture that can compress/decompress 64 bytes/cycle with minor overhead. Moreover, it provides high compression ratio on small block sizes as well as large ones.
Addresses the problem of accessing portions of multiband data which has been losslessly compressed. An approach that uses the fractal property of some well known space filling curves to provide access to portions of a...
详细信息
Addresses the problem of accessing portions of multiband data which has been losslessly compressed. An approach that uses the fractal property of some well known space filling curves to provide access to portions of a losslessly compressed data set is described. This approach reduces the average amount of decompression necessary to access any portion of the data set, thereby reducing the amount of time required to access the compressed data. Various tradeoffs exist which will be discussed with practical examples.< >
暂无评论