A system comprising a head worn display and an image rendering computing device is disclosed. The head worn display comprises one or more image sensors, a head position sensor, and a head position computing device configured to determine a first head pose. The image rendering computing device is configured to render an enhanced image and embed the first head pose in the enhanced image. The head position computing platform is further configured to determine a second head pose, determine a difference between the second head pose and the first head pose, and warp the enhanced image based on the difference between the second head pose and the first head pose to correct for head motion of the user.

A building appraisal system conducted by a remote inspector located away from a building. A remote user connected to a user on site can share images, measurements and other data to conduct an examination of the building such as an appraisal. A processor coupled to an image sensor can be configured to receive a gross floor area of the building. Images of an interior room of the building are stored in memory. The processor determines a planar surface in the images corresponding to a floor surface of the interior room and a plurality of corners in the images forming vertices of a bounded floor area on the floor surface. The processor can compute an adjusted floor area of the building that includes the bounded floor area subtracted from the gross floor area.

In accordance with some embodiments of the disclosed subject matter, methods, systems, and media for generating images of multiple sides of an object are provided. In some embodiments, a method comprises receiving information indicative of a 3D pose of a first object in a first coordinate space at a first time; receiving a group of images captured using at least one image sensor, each image associated with a field of view within the first coordinate space; mapping at least a portion of a surface of the first object to a 2D area with respect to the image based on the 3D pose of the first object; associating, for images including the surface, a portion of that image with the surface of the first object based on the 2D area; and generating a composite image of the surface using images associated with the surface.

A dual variational autoencoder-generative adversarial network (VAE-GAN) is trained to transform a real video sequence and a simulated video sequence by inputting the real video data into a real video decoder and a real video encoder and inputting the simulated video data into a synthetic video encoder and a synthetic video decoder. Real loss functions and simulated loss functions are determined based on output from a real video discriminator and a simulated video discriminator, respectively. The real loss functions are backpropagated through the real video encoder and the real video decoder to train the real video encoder and the real video decoder based on the real loss functions. The synthetic loss functions are backpropagated through the synthetic video encoder and the synthetic video decoder to train the synthetic video encoder and the synthetic video decoder based on the synthetic loss functions. The real video discriminator and the synthetic video discriminator can be trained to determine an authentic video sequence from a fake video sequence using the real loss functions and the synthetic loss functions. The annotated simulated video can be transformed with the synthetic video encoder and the real video decoder of the dual VAE-GAN to generate a reconstructed annotated real video sequence that includes style elements based on the real video sequence. A second neural network is trained using the reconstructed annotated real video sequence to detect and track objects.

A dual variational autoencoder-generative adversarial network (VAE-GAN) is trained to transform a real video sequence and a simulated video sequence by inputting the real video data into a real video decoder and a real video encoder and inputting the simulated video data into a synthetic video encoder and a synthetic video decoder. Real loss functions and simulated loss functions are determined based on output from a real video discriminator and a simulated video discriminator, respectively. The real loss functions are backpropagated through the real video encoder and the real video decoder to train the real video encoder and the real video decoder based on the real loss functions. The synthetic loss functions are backpropagated through the synthetic video encoder and the synthetic video decoder to train the synthetic video encoder and the synthetic video decoder based on the synthetic loss functions. The real video discriminator and the synthetic video discriminator can be trained to determine an authentic video sequence from a fake video sequence using the real loss functions and the synthetic loss functions. The annotated simulated video can be transformed with the synthetic video encoder and the real video decoder of the dual VAE-GAN to generate a reconstructed annotated real video sequence that includes style elements based on the real video sequence. A second neural network is trained using the reconstructed annotated real video sequence to detect and track objects.

There is provided an image processing apparatus comprising: a processor; and a memory storing a program. When the program is executed by the processor, the program causes the image processing apparatus to: obtain a first RAW image including a region of a first circular fisheye image; and develop the first RAW image. A pixel outside the region of the first circular fisheye image in the first RAW image is not developed.

One embodiment provides a method, including: identifying, using data obtained from at least one sensor associated with an information handling device, a multi-planar orientation of a non-flat display surface of the information handling device and a spatial orientation of the information handling device with respect to a user's gaze position; determining, using a processor and based on the identifying, a distortion of at least one object displayed on the non-flat display surface; and adjusting at least one aspect of the non-flat display surface to correct the distortion. Other aspects are described and claimed.

Embodiments of the present invention provide systems and methods to detect a moving anatomic feature during a treatment sequence based on a computed and/or a measured shortest distance between the anatomic feature and at least a portion of an imaging system.

An apparatus for stitching together multiple camera images to form a blended image having an output projection format. The apparatus is configured to convert each of the multiple camera images into the output projection format. It is configured to stitch together the converted images to form a single image. It is also configured to output the single image as the blended image having the output projection format.

An improved method for coding video is provided that includes Virtual Reality (VR) sequences that enables more efficient encoding by organizing the VR sequence as a single 2D block structure. In the method, reference picture and subpicture lists are created and extended to account for coding of the VR sequence. To further improve coding efficiency, reference indexing can be provided for the temporal and spatial difference between a current VR picture block and the reference pictures and subpictures for the VR sequence. Further, because the reference subpictures for the VR sequence may not have the proper orientation once the VR sequence subpictures are organized into the VR sequence, reorientation of the reference subpictures is made so that the reference subpicture orientations match the current VR subpicture orientations.