International Business Machines Corporation

Huawei Technologies Co., Limited

Lattice Semiconductor Corporation

Cerence Operating Company

Yi Gao

Long Wen Lan

Xing Liu

Kai Zhou

Henry Esmond Butterworth

Jia Wang

Qi Zhou

Li Hui Guo

Yao Zhou

Kai Lei

Qiming Wu

Weiran Nie

Jianzhong Teng

Xiao-Lin Ren

Wen Wu Na

Jiandong Ke

Youjia Huang

Bao Sheng Lu

Jia Jun Liu

Xingui Zeng

Xiaolin Ren

A speech detection method includes performing recognition on a photographed face image using a model to predict whether a user intends to continue speaking, and to determine whether a collected audio signal is a speech end point with reference to a prediction result.

Speech detection method, prediction model training method, apparatus, device, and medium

Methods and an apparatus for performing feature extraction on speech in a microphone signal with embedded noise processing to reduce the amount of processing are provided. In embodiments, feature extraction and the noise estimate use an output of the same Fourier Transform, such that the noise filtering of the speech is embedded with the feature extraction of the speech.

Methods and apparatus for ASR with embedded noise reduction

Provided is a method, system, and computer program product for managing requests received by a storage system. The method may include detecting, by one or more processors, a failure in a first storage system in response to a request to access data in the first storage system. The first storage system may include a primary storage for storing the data. A second storage system may be activated in response to the detected failure. The second storage system may include a first storage and a second storage. The first storage may include data synchronized with the primary storage of the first storage system. The second storage may be used to store data that corresponds to the request. The request may be managed using the second storage system.

Storage system management

Systems, computer program products, and methods that can integrate resources at a disaster recovery site are provided. One method includes generating, on a primary site, a set of storage snapshots based on combining another set of storage snapshots and incremental changes to mirrored data on a backup site in which the storage snapshots include a second snapshot format utilized on the backup site. The method further includes converting the set of storage snapshots from the second snapshot format to a first snapshot format utilized on the primary site to generate a yet another set of storage snapshots and converting, on the backup site, a set of storage snapshots including the incremental changes and the second snapshot format to the first snapshot format to generate still another set of eighth storage snapshots. The storage snapshots on both sites represent the same data in the same snapshot format without bulk data transfer.

Integrating resources at a backup site

Embodiments relate to enhancing echo cancellation in a transceiver integrated circuit (IC) for full-duplex communication by providing a signal path connected to a dummy driver that replicates a signal path between a main driver and a counterpart transceiver IC to cause a duplicated signal generated by the dummy driver to more closely replicate a sending signal generated by the main driver. The signal path connected to the dummy driver includes circuit elements having transmission line parameters and RLC parameters that replicate transmission line parameters and RLC parameters of circuit elements in the signal path between the main driver and the counterpart transceiver IC.

Enhanced echo cancellation in full-duplex communication

A video stream encoding apparatus and method therefore is herein described. The video stream encoding apparatus includes a divider for dividing a frame in a video stream into a plurality of processing units, each processing unit including a plurality of macro-blocks; and an encoder for encoding the plurality of macro-blocks in each processing unit in parallel. The same quantization parameters are used for respective macro-blocks in the same processing unit in the quantization of the encoder. By the apparatus and the method of the invention, dependencies between macro-blocks are broken, and the precision of high rate controlling is kept while encoding the respective macro-blocks in parallel.

Macro-block video stream encoding

Growing community of inventors

Shanghai, China

Yi Gao