Black Sesame Technologies Inc.

Tiecheng Wu

Bo Li

Jin Xu

Jizhang Shan

Xibeijia Guan

Fangwen Tu

The invention discloses a system for image segmentation for detecting an object. More particularly, the system performs two-stage segmentation on the object in the image to generate an enhanced image. The first stage is the object detection, followed by a second stage including segmentation. The invention segments the object from a background of the image to create an enhanced image.

Strip decoder and inside-outside loss for high quality image segmentation

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training a machine learning model configured to generate a predicted depth image, comprising receiving data representing training samples that include a plurality of image pairs, each image pair includes a target image and a reference image both capturing a particular scene from different orientations; for each of the plurality of image pairs, generating a compressed cost volume for the image pair; providing the compressed cost volume as an input to the machine learning model; generating, using the machine learning model, output data representing a predicted disparity map for the compressed cost volume; and generating a total loss using the predicted disparity map for the compressed cost volume, the total loss includes a boundary loss, an occlusion loss, and a transfer loss; and updating the plurality of parameters of the machine learning model by minimizing the total losses.

Generating depth images for image data

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating a depth image, comprising obtaining data representing a first image generated by a first sensor and a second image generated by a second sensor, wherein each of the first and second images includes a plurality of pixels; determining, for each pixel of the plurality of pixels included in the first image, whether the pixel is a boundary pixel associated with a boundary of an object that is represented in the first image; determining, from a plurality of candidate penalty values and for each pixel in the first image, an optimized penalty value for the pixel; generating an optimized cost function for the first image based on the optimized penalty values for the plurality of pixels; and generating a depth image for the first image based on the optimized cost function.

Generating stereo-based dense depth images

A method of motion control and textured display, comprising, receiving an image, extracting an initial mask of a subject of interest from the image, segmenting the image into the initial mask and a background, determining a color texture of the initial mask, estimating motion of the initial mask to a new mask and merging the new mask and the background into a merged image.

Subject segmentation for motion control and textured display

A method of depth-of-field simulation, including receiving a plurality of images, predicting a layer of interest mask of the plurality of images, determining a plurality of mean brightness anchor values of a respective plurality of layers of interest window arrays, setting a plurality of layers of interest set of binary codes, determining a hamming distance between plurality of layers of interest set of binary codes, determining a cost volume based on the hamming distance, resampling a vertical cost based on a vertical ordinal direction the cost volume, resampling a horizontal cost based on a horizontal ordinal direction the cost volume, determining an all-in-focus layer based on the vertical cost and the horizontal cost, determining an out-of-focus layer based on the vertical cost and the horizontal cost and determining a depth of the all-in-focus layer and out-of-focus layer.

Depth-of-field simulation via optical-based depth extraction

The present invention discloses system and method for replacing a texture of a background region in a multimedia. The system is an AI-based multi-media processing system for replacing the original texture of the background of the multimedia with a texture template. The system applies a foreground mask to hide and protect the foreground region and multiple textures of the background. The system uses deep learning to segment specific texture from images or video sequences. The system replaces the texture of the original input image with a texture template to form a processed image.

Growing community of inventors

Singapore, Singapore

Tiecheng Wu