Tencent America LLC

Shih-Yao Lin

Wei Fan

Hui Tang

Lianyi Han

Chao Huang

Yusheng Xie

Zhimin Huo

Tao Yang

Kun Wang

Qian Zhen

Zhen Qian

Min Tu

Nan Du

Xiaozhong Chen

A method, computer program, and computer system are provided for image segmentation. Image data, such as biological image data, is received. One or more objects associated with the received image data is detected. One or more regions of interest are determined within the receive image data corresponding to one or more segments based on the detected objects.

System and method for instance segmentation in pathological images using object-guided framework

There is included a method and apparatus comprising computer code configured to cause a processor or processors to perform obtaining video data including at least one body part of a person, selecting keypoints of the at least one body part based on a predetermined rehabilitation category, extracting a motion feature of the at least one body part from the video data, scoring the motion feature based on the predetermined rehabilitation category, and generating a display illustrating the motion feature and said scoring of the motion feature.

Vision-based motion capture system for rehabilitation training

A method and user device for determining a unified Parkinson's disease rating scale (UPDRS) value associated with a user of the user device include obtaining video data associated with a movement of a body part of the user. The UPDRS value is determined using a model and the video data associated with the movement of the body part of the user. The UPDRS value is provided to permit an evaluation of the user based on the UPDRS value.

User device based Parkinson's disease detection

Systems and methods for obtaining hand images are provided. A method, performed by at least one processor that implements at least one network, includes obtaining a single source image, a three-dimensional (3D) hand pose of a first hand in the single source image, and a 3D target hand pose; and generating an image of a second hand, that has an appearance of the first hand and a pose of the 3D target hand pose, based on the single source image, the 3D hand pose, and the 3D target hand pose.

Synthesizing 3D hand pose based on multi-modal guided generative networks

A method and user device for providing a unified Parkinson's disease rating scale (UPDRS) value by a user device includes generating, by a sensor of the user device, sensor data based on a user of the user device manipulating the user device. The UPDRS value is determined based on the sensor data and a deep neural network (DNN) model. The UPDRS value is provided to permit an evaluation of the user based on the UPDRS value.

User device based parkinson disease detection

There is included an apparatus and system including video capture code configured to control a camera of a mobile phone to capture a video of at least a portion of skin, extraction code configured to extract at least one photoplethysmogram (PPG) signal from the video, determination code configured to determine a peripheral oxygen saturation (SpO2) value based on the PPG signal with respect to a blood flow at the portion of skin, and display code configured to control a display of the mobile phone to display the SpO2 value.

SpO2 miniprogram: AI SpO2 measurement app

A method and apparatus are provided that includes iteratively sampling candidates from medical records and evaluating whether ones of the candidates better explain a member from the medical records. The iterations replace the member with the candidates and depending on whether the candidates better explain the member from the medical records may be weighted in a next iteration.

Method for determining disease symptom relations using acceptance and rejection of random samples

A computer-implemented method of training a neural network for organ segmentation may be provided. The method may include: collecting a set of digital sample images from a database; inputting the collected set of digital images into a neural network recognition model; and training the neural network recognition model to recognize a first object in a first digital image as a specific object based on the first object being similar to a second object in a second digital image. The method may comprise predicting a signed distance map (SDM) in conjunction with a segmentation map.

Shape-aware organ segmentation by predicting signed distance maps

A method, computer program, and computer system is provided for estimating three-dimensional hand poses in images. Data corresponding to two hand images is receive, and an optical flow value corresponding to a change in a hand gesture in the received hand image data is calculate. A heat map is generated based on the calculated optical flow, and a hand mesh map is estimated based on the generated heat map. A hand pose present within the hand images is determined based on the estimated hand mesh map.

Video-based 3D hand pose and mesh estimation based on temporal-aware self-supervised learning

A computer-implemented method, computer readable storage medium, and computer system is provided for estimating three-dimensional (3D) hand poses in images by receiving data corresponding to a hand image, generating a depth map corresponding to the received hand image data, and estimating a hand pose from the received hand image data and the generated depth map.

3D hand pose estimation based on depth-image guided adversarial network 1

A method and apparatus for generating augmented training data for hand pose estimation include receiving source data that is associated with a first lighting condition. Target data that is associated with a second lighting condition is received. A lighting condition translation between the first lighting condition and the second lighting condition is determined. Lighting translated data is generated based on the lighting condition translation and the source data. Augmented training data for hand pose estimation is generated based on the target data and the lighting translated data.

Augmenting reliable training data with CycleGAN for hand pose estimation

A method of synthesizing an image of a hand using a blending generative adversarial network (BlendGAN) includes obtaining a synthetic 3-dimensional (3D) hand pose including a 3D model of a hand; obtaining a real background image; combining the synthetic 3D hand pose with the real background image to create a synthetic hand image; and blending the synthetic hand image using the BlendGAN to create a blended synthetic hand image.

Growing community of inventors

Palo Alto, CA, United States of America

Shih-Yao Lin