No abstract

The past decade has witnessed the great success of deep learning in many disciplines, especially in computer vision and image processing. However, deep learning-based video coding remains in its infancy. We review the representative works about using deep learning for image/video coding, an actively developing research area since 2015. We divide the related works into two categories: new coding schemes that are built primarily upon deep networks, and deep network-based coding tools that shall be used within traditional coding schemes. For deep schemes, pixel probability modeling and auto-encoder are the two approaches, that can be viewed as predictive coding and transform coding, respectively. For deep tools, there have been several techniques using deep learning to perform intra-picture prediction, inter-picture prediction, cross-channel prediction, probability distribution prediction, transform, post- or in-loop filtering, down- and up-sampling, as well as encoding optimizations. In the hope of advocating the research of deep learning-based video coding, we present a case study of our developed prototype video codec, Deep Learning Video Coding (DLVC). DLVC features two deep tools that are both based on convolutional neural network (CNN), namely CNN-based in-loop filter and CNN-based block adaptive resolution coding. The source code of DLVC has been released for future research.

In this paper, we propose a novel patient-specific electro cardiogram (ECG) classification algorithm based on the recurrent neural networks (RNN) and density based clus tering technique. We use RNN to learn time correlation among ECG signal points and to classify ECG beats with different heart rates. Morphology information including the present beat and the T wave of former beat is fed into RNN to learn underlying features automatically. Clus tering method is employed to find representative beats as the training data. Evaluated on the MIT-BIB Arrhythmia Database, the experimental results show that proposed al gorithm achieves the state-of-the-art classification perfor mance.

No abstract

In this proposal, we design a learned multifrequency image compression approach that uses generalized octave convolutions to factorize the latent representations into high-frequency (HF) and low-frequency (LF) components, and the LF components have lower resolution than HF components, which can improve the rate-distortion performance, similar to wavelet transform. Moreover, compared to the original octave convolution, the proposed generalized octave convolution (GoConv) and octave transposed-convolution (GoTConv) with internal activation layers preserve more spatial structure of the information, and enable more effective filtering between the HF and LF components, which further improve the performance. In addition, we develop a variable-rate scheme using the Lagrangian parameter to modulate all the internal feature maps in the autoencoder, which allows the scheme to achieve the large bitrate range of the JPEG AI with only three models. Experiments show that the proposed scheme achieves much better Y MS-SSIM than VVC. In terms of YUV PSNR, our scheme is very similar to HEVC.Index Terms-learned image compression, octave convolution, variable-rate deep learning models, modulated scheme I. METHOD A. Overview of the Encoding/Decoding Architectures Recently, deep learning-based image compression has shown the potential to outperform standard codecs such as JPEG2000 [7], the H.265/HEVC Intra-based BPG image codec [8], and the intra coding of the upcoming versatile video coding (VVC) standard [5]. In particular, the scheme in [1] even achieves better PSNR than VVC for the Kodak