Tencent Technology (shenzhen) Company Limited

Tencent Technology Company Limited

Jingmin Luo

Xiaolong Zhu

Yitong Wang

Lijian Mei

Shenghui Huang

Kaining Huang

Haozhi Huang

Yongsen Zheng

Yan Zhang

Yi Tong Wang

Xing Ji

Liang Qiao

Kai Ning Huang

Xiao Long Zhu

Xinyu Gong

This application discloses a foreground data generation method performed at a computer device. The method includes: obtaining a target image, the target image containing a target object and a background; removing the background from the target image to obtain initial foreground data of the target object in the target image; expanding the initial foreground data and eroding the expanded foreground data; blurring the eroded foreground data to obtain certain foreground data and uncertain data from the initial foreground data; and segmenting the certain foreground data from the uncertain data, to obtain target foreground data of the target object in the target image.

Foreground data generation method and method for applying same, related apparatus, and system

Embodiments of this application provide a neural network model deployment method, a prediction method and a device. The described features can implement deployment of a neural network model to improve the universality of the deployment of the neural network model to the terminal device by obtaining a layer definition and an operation parameter of each network layer of an initial neural network model, executing a target network layer corresponding to the network layers, applying relational connections amongst the target network layers using a net class, converting the operation parameters into a preset format, obtaining a target operation parameter based on the preset format, loading a corresponding target operation parameter in the target network layer, and obtaining a target neural network model based on the target operation parameter.

Neural network model deployment method, prediction method and related device

Embodiments of this application disclose a foreground data generation method performed at a computer device. The method includes: obtaining a background image and a target image, the target image containing a target object and a background; removing the background from the target image according to the background image and the target image, to obtain initial foreground data of the target object in the target image; obtaining certain foreground data and uncertain data from the initial foreground data, wherein the uncertain data represents data whose value is between the certain foreground data and background data corresponding to the background; and segmenting the certain foreground data from the uncertain data, to obtain target foreground data of the target object in the target image.

This application provides a method for training a pose recognition model performed at a computer device. The method includes: inputting a sample image labeled with human body key points into a feature map model included in a pose recognition model, to output a feature map of the sample image; inputting the feature map into a two-dimensional (2D) model included in the pose recognition model, to output 2D key point parameters used for representing a 2D human body pose; input a target human body feature map cropped from the feature map and the 2D key point parameter into a three-dimensional (3D) model included in the pose recognition model, to output 3D pose parameters used for representing a 3D human body pose; constructing a target loss function based on the 2D key point parameters and the 3D pose parameters; and updating the pose recognition model based on the target loss function.

Method and apparatus for training pose recognition model, and method and apparatus for image recognition

A computer device extracts a plurality of target windows from a target video. Each of the target windows comprises a respective plurality of consecutive video frames. For each of the target windows, the device performs action recognition on the respective plurality of consecutive video frames corresponding to the target window to obtain respective first action feature information of the target window. The device obtains a similarity between the first action feature information of the target window and preset feature information. The device determines, from the respective obtained similarities corresponding to the plurality of target windows, a highest first similarity and a first target window corresponding to the highest first similarity. The device also determines a dynamic action corresponding to the highest first similarity as the preset dynamic action in accordance with threshold settings.

Action recognition method and apparatus, and human-machine interaction method and apparatus

In a multi-person pose recognition method, a to-be-recognized image is obtained, and a circuitous pyramid network is constructed. The circuitous network pyramid includes parallel phases, and each phase includes downsampling network layers, upsampling network layers, and a first residual connection layer to connect the downsampling and upsampling network layers. The phases are interconnected by a second residual connection layer. The circuitous pyramid network is traversed, by extracting a feature map for each phase, and the feature map of the last phase is determined to be the feature map of the to-be-recognized image. Multi-pose recognition is then performed on the to-be-recognized image according to the feature map to obtain a pose recognition result for the to-be-recognized image.

Multi-person pose recognition method and apparatus, electronic device, and storage medium

An image recognition method is provided. The method includes obtaining predicted locations of joints of a target person in a to-be-recognized image based on a joint prediction model, where the joint prediction model is pre-constructed by: obtaining a plurality of sample images; inputting training features of the sample images and a body model feature to a neural network and obtaining predicted locations of joints in the sample images outputted by the neural network; updating a body extraction parameter and an alignment parameter; and inputting the training features of the sample images and the body model feature to the neural network to obtain the joint prediction model.

Image recognition method and apparatus, electronic device, and readable storage medium using an update on body extraction parameter and alignment parameter

An image processing method and a related apparatus are provided. The method is applied to an image processing device, and includes: obtaining an original image, the original image including a foreground object; extracting a foreground region from the original image through a deep neural network; identifying pixels of the foreground object from the foreground region; forming a mask according to the pixels of the foreground object, the mask including mask values corresponding to the pixels of the foreground object; and extracting the foreground object from the original image according to the mask.

Image processing method and apparatus

An augmented reality processing method is provided for a terminal. The method includes: obtaining a plurality of frames of images, comprising a first image and a second image, which is a frame of an image immediately following the first image; obtaining a key point set of a first object in the first image; obtaining, through a neural network model, first pose key point sets respectively corresponding to a plurality of objects in the second image; determining a second pose key point set of the first object in the second image according to the key point set and a motion trend of the first object; using a target first pose key point set as a key point set of the first object in the second image; and generating an augmented information image according to the key point set of the first object in the second image.

Augmented reality processing method, object recognition method, and related device

A video image processing method and apparatus is described. The method includes obtaining a first key point of an Mframe image in a video. The method further includes performing weighted smoothing on the first key point of the Mframe image according to first key points in a historical key point queue and weights in a first target weight queue, to obtain a target key point. The historical key point queue includes a first key point corresponds to each frame image in N frame images. The N frame images are images before the Mframe image, N&#x3e;0. The weights in the first target weight queue corresponding to the first key points in the historical key point queue, A weight corresponding to a first key point of an (M-a)frame image being greater than or equal to a weight corresponding to a first key point of an (M-b)frame image, and a&#x3c;b. The method further includes adjusting the Mframe image according to the target key point.

Growing community of inventors

Shenzhen, China

Jingmin Luo