Self Electronics Co., Ltd.

Alipay (hangzhou) Information Technology Co., Ltd.

Alibaba Group Holding Limited

Advanced New Technologies Co., Ltd.

Zhiming Wang

Jun Zhou

Xiaolong Li

Kaisheng Yao

Bozhang Xu

Zhaoyong Zheng

Li Wang

Chengke Zhang

Junjun Ying

Xuhong Ma

Xin Long

Liang You

Niqi Lyu

Provided are a video stream processing method and apparatus, an electronic device and a computer-readable medium, which relate to the field of Internet technology. An exemplary method includes: storing a video frame in an original video stream in a first buffer area by calling a video stream processing interface of a video stream processing tool; processing the video frame in the first buffer area by means of a video frame processing model to obtain a processed video frame; and on the basis of the processed video frame, generating a standard video stream corresponding to the original video stream. By means of the embodiments of the this application, the software complexity of video stream processing is effectively reduced, and the speed of video stream processing is effectively improved.

Video stream processing method and apparatus, and electronic device and computer-readable medium

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for determining voice characteristics are provided. One of the methods includes: obtaining speech data of a speaker; inputting the speech data into a model trained at least by jointly minimizing a first loss function and a second loss function, wherein the first loss function comprises a non-sampling-based loss function and the second loss function comprises a Gaussian mixture loss function with non-unit multi-variant covariance matrix; and obtaining from the trained model one or more voice characteristics of the speaker.

System and method for determining voice characteristics

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for personalized speaker verification are provided. One of the methods includes: obtaining first speech data of a speaker as a positive sample and second speech data of an entity different from the speaker as a negative sample; feeding the positive sample and the negative sample to a first model for determining voice characteristics to correspondingly output a positive voice characteristic and a negative voice characteristic of the speaker; obtaining a gradient based at least on the positive voice characteristic and the negative voice characteristic; and feeding the gradient to the first model to update one or more parameters of the first model to obtain a second model for personalized speaker verification.

System and method for personalized speaker verification

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for determining voice characteristics are provided. One of the methods includes: obtaining speech data of a speaker; inputting the speech data into a model trained at least by jointly minimizing a first loss function and a second loss function, wherein the first loss function comprises a non-sampling-based loss function and the second function comprises a Gaussian mixture loss function with non-unit multi-variant covariance matrix; and obtaining from the trained model one or more voice characteristics of the speaker.

A speech wakeup method, apparatus, and electronic device are disclosed in embodiments of this specification. The method includes: inputting speech data to a speech wakeup model trained with general speech data; and outputting, by the speech wakeup model, a result for determining whether to execute speech wakeup, wherein the speech wakeup model includes a Deep Neural Network (DNN) and a Connectionist Temporal Classifier (CTC).

Speech wakeup method, apparatus, and electronic device

A non-polar rectifying circuit includes two input terminals, two output terminals, two P-channel MOS transistors, and two N-channel MOS transistors. The two input terminals are respectively connected with a drain of one of the two P-channel MOS transistors and a drain of one of the two N-channel MOS transistors. One of the output terminals is electrically connected with the source of two P-channel MOS transistors, and the other output terminal is electrically connected with the source of two N-channel MOS transistors. In application, regardless of the two input terminals which is the positive polarity or negative polarity, one of the output terminals will output high level, and another will also output low level. As a result, it make the power plug no need to divide the positive and negative, which can reduce the installation time of LED lamps, especially for installing a large number of LED lightings.

Non-polar rectifying circuit

An interconnection mechanism for LED bar lighting includes a plug mechanism and a socket mechanism. The plug mechanism includes a hollow socket, a plug, and a connection pin. The hollow socket includes a hollow plug barrel, a stop ring, a wedge importing ring, and at least one opening. A slope of the wedge importing ring extends toward a free end of the hollow plug barrel. The connection pin includes at least one plug locating ring, at least one socket locating ring, and at least one socket locating bar. The plug locating ring is arranged between the stop ring and the wedge importing ring for fixing the relative position of the connection pin and the hollow socket. The socket mechanism includes a hollow socket barrel, and a locating slot. The socket locating bar is engaged with the locating slot so as to fix the relative position of the connection pin and the hollow socket barrel.

Interconnection mechanism for LED bar lighting

An LED strip-shaped lamp with a connecting structure includes a first lamp housing, a plug disposed on the first lamp housing, a second lamp housing coupled with the first lamp housing, and a socket on the second lamp housing coupled to the plug. The plug includes a first magnet disposed at an end of the first lamp housing and a first electrical connector fixed in the first pad. The socket includes a second magnet at an end of the second lamp housing and a second electrical connector fixed in the second pad. When the magnets are coupled, the first and the second electrical connector are coupled to connect electrically. The magnets used in the connecting structure of the LED strip-shaped lamp of the present invention can be used as a structure for connecting the first and second lamp housing so as to simplify the connection of the strip-shaped lamps.

LED strip-shaped lamp with a connecting structure

A power supply system for shelf track lighting device includes a track, and a jointer. The track includes a U-shaped slot, two grooves, and two conductive posts. The jointer includes a bottom cover, an upper cover, and two elastic chips. The upper cover includes a T-shaped supporting cover. The two elastic chips are sandwiched between the T-shaped supporting cover and the bottom cover. The elastic chip includes an arm, a supporting end, an arc-shaped bend portion, and three right-angle-shaped bend portion. The arc-shaped bend portion is near the arm and the arm holds out of the T-shaped supporting cover and abuts against the conductive post. The arm is retracted into the T-shaped supporting cover under the resilient deformation force of the three right-angle-shaped bend portions, thereby tightly connecting the elastic chip to the conductive post is achieved, and further, power can be supplied to the strip LED lamp.

Growing community of inventors

Zhejiang, China

Zhiming Wang