Described herein are apparatuses, systems and methods for generating an interactive three-dimensional (“3D”) environment using virtual depth. A method comprises receiving a pre-rendered media file comprising a plurality of frames, receiving depth data related to the media file, wherein the depth data corresponds to each of the plurality of frames, creating an invisible three-dimensional (“3D”) framework of a first frame of the media file based on the corresponding depth data, and rendering a new first frame in real time to include the pre-rendered first frame, one or more virtual visible 3D objects and the invisible 3D framework.

Techniques for managing resources among clusters of computing devices in a computing system are described herein. In one embodiment, a method includes receiving, via a computer network, a resource reassignment message indicating that a server is reassigned from a first cluster to a second cluster and in response to the received resource reassignment message, establishing communications with the server reassigned from the first cluster to the second cluster via the computer network. The method further includes subsequent to establishing communications with the server via the computer network, assigning a compute load to the server reassigned from the first cluster to the second cluster without physically relocating the server from the first cluster to the second cluster.

Embodiments of the present disclosure disclose a method and an apparatus for operating a control system, a storage medium, and an electronic apparatus. The method includes: obtaining, by using a target recording device in a control system, first image information of a target object moving in a real scene; obtaining a first distance corresponding to the first image information, the first distance being a distance between the target recording device and the target object; and adjusting a target parameter of the control system according to the first distance, the target parameter being used for outputting media information to a virtual reality (VR) device, the VR device being connected to the control system, the media information being corresponding to movement information of the target object moving in the real scene, and the movement information comprising the first distance.

A parallax correction method and device, and a storage medium. The method comprises: collecting two original images comprising a target object by means of a binocular camera; determining a first paralax of the target object in imaging areas of the two original images; adjusting the positions of the imaging areas in the two original images according to the first parallax and a preset parallax; and determining a target image on the basis of the imaging areas after position adjustment.

A multi-aperture imaging device includes image sensor means with a plurality of image sensor areas and a plurality of optical channels, wherein each optical channel includes an optic for imaging a partial field of view of a total field of view onto an image sensor area of the image sensor means associated with the optical channel. The plurality of optical channels is configured to image the total field of view completely. A first partial field of view of the total field of view and a second partial field of view of the total field of view are captured by a different number of optical channels.

The image processing apparatus that generates a virtual viewpoint image includes a generation unit. The generation unit generates a virtual viewpoint image in which in accordance with a virtual viewpoint specified by the virtual viewpoint information. The degree of transparency of a specific object at a position a specific distance apart from a position of a first virtual viewpoint in a virtual viewpoint image in accordance with the first virtual viewpoint corresponding to a first angle of view, which is generated by the generation unit, is higher than a degree of transparency of an object at a position the specific distance apart from a position of a second virtual viewpoint in a virtual viewpoint image in accordance with the second virtual viewpoint corresponding to a second angle of view larger than the first angle of view.

The image processing apparatus that generates a virtual viewpoint image includes a generation unit. The generation unit generates a virtual viewpoint image in which in accordance with a virtual viewpoint specified by the virtual viewpoint information. The degree of transparency of a specific object at a position a specific distance apart from a position of a first virtual viewpoint in a virtual viewpoint image in accordance with the first virtual viewpoint corresponding to a first angle of view, which is generated by the generation unit, is higher than a degree of transparency of an object at a position the specific distance apart from a position of a second virtual viewpoint in a virtual viewpoint image in accordance with the second virtual viewpoint corresponding to a second angle of view larger than the first angle of view.

A camera module according to the present embodiment includes a light emitting part configured to output a light signal to an object, a filter configured to allow the light signal to pass therethrough, at least one lens disposed on the filter and configured to collect the light signal from the object, a sensor configured to generate an electric signal from the light signal collected by the lens, the sensor including a plurality of pixels arranged in an array form, and a tilting part configured to tilt the filter to repeatedly move an optical path of the light signal having passed through the filter according to a predetermined rule. The optical path of the light signal passing through the filter is moved in one direction among diagonal directions of the sensor with respect to an optical path corresponding to the filter being disposed parallel to the sensor.

The disclosed subject matter is directed to employing machine learning models configured to predict 3D data from 2D images using deep learning techniques to derive 3D data for the 2D images. In some embodiments, a method is provided that comprises receiving, by a system comprising a processor, a panoramic image, and employing, by the system, a three-dimensional data from two-dimensional data (3D-from-2D) convolutional neural network model to derive three-dimensional data from the panoramic image, wherein the 3D-from-2D convolutional neural network model employs convolutional layers that wrap around the panoramic image as projected on a two-dimensional plane to facilitate deriving the three-dimensional data.

The disclosed subject matter is directed to employing machine learning models configured to predict 3D data from 2D images using deep learning techniques to derive 3D data for the 2D images. In some embodiments, a method is provided that comprises receiving, by a system comprising a processor, a panoramic image, and employing, by the system, a three-dimensional data from two-dimensional data (3D-from-2D) convolutional neural network model to derive three-dimensional data from the panoramic image, wherein the 3D-from-2D convolutional neural network model employs convolutional layers that wrap around the panoramic image as projected on a two-dimensional plane to facilitate deriving the three-dimensional data.