Abstract: In one embodiment, a method includes generating, by a device, first tracking data using a first tracking algorithm, based on first video frames associated with a scene. An augmented-reality (AR) effect may be displayed based on the first tracking data. The device may generate a first confidence score associated with the first tracking data and determine that the first confidence score is above a threshold. The device may generate, based on second video frames subsequent to the first video frames, second tracking data using the first tracking algorithm. The device may determine that an associated second confidence score is below a threshold. In response, the device may generate, based on third video frames subsequent to the second video frames, third tracking data using a second tracking algorithm different from the first. The device may then display the AR effect based on the third tracking data.

Abstract: In one embodiment, a computing system may access a training image and a reference image of a person and an incomplete image. A generate may generate an in-painted image based on the incomplete image, and a discriminator may be used to determine whether each of the in-painted image, the training image, and the reference image is likely generated by the generator. The system may compute losses based on the determinations and update the discriminator accordingly. Using the updated discriminator, the system may determine whether a second in-painted image generated by the generator is likely generated by the generator. The system may compute a loss based on the determination and update the generator accordingly. Once training is complete, the generator may be used to generate a modified version of a given image, such as making the eyes of a person appear open even if they were closed in the input image.

Abstract: In one embodiment, a computing system may access a training image and a reference image of a person and an incomplete image. A generate may generate an in-painted image based on the incomplete image, and a discriminator may be used to determine whether each of the in-painted image, the training image, and the reference image is likely generated by the generator. The system may compute losses based on the determinations and update the discriminator accordingly. Using the updated discriminator, the system may determine whether a second in-painted image generated by the generator is likely generated by the generator. The system may compute a loss based on the determination and update the generator accordingly. Once training is complete, the generator may be used to generate a modified version of a given image, such as making the eyes of a person appear open even if they were closed in the input image.

Abstract: In one embodiment, a method includes generating, by a device, first tracking data using a first tracking algorithm, based on first video frames associated with a scene. An augmented-reality (AR) effect may be displayed based on the first tracking data. The device may generate a first confidence score associated with the first tracking data and determine that the first confidence score is above a threshold. The device may generate, based on second video frames subsequent to the first video frames, second tracking data using the first tracking algorithm. The device may determine that an associated second confidence score is below a threshold. In response, the device may generate, based on third video frames subsequent to the second video frames, third tracking data using a second tracking algorithm different from the first. The device may then display the AR effect based on the third tracking data.