FYUSION, INC.

Various embodiments of the present invention relate generally to systems and processes for transforming a style of video data. In one embodiment, a neural network is used to interpolate native video data received from a camera system on a mobile device in real-time. The interpolation converts the live native video data into a particular style. For example, the style can be associated with a particular artist or a particular theme. The stylized video data can viewed on a display of the mobile device in a manner similar to which native live video data is output to the display. Thus, the stylized video data, which is viewed on the display, is consistent with a current position and orientation of the camera system on the display.

Provided are mechanisms and processes for performing visual search using multi-view digital media representations, such as surround views. In one example, a process includes receiving a visual search query that includes a surround view of an object to be searched, where the surround view includes spatial information, scale information, and different viewpoint images of the object. The surround view is compared to stored surround views by comparing spatial information and scale information of the surround view to spatial information and scale information of the stored surround views. A correspondence measure is then generated indicating the degree of similarity between the surround view and a possible match. At least one search result is then transmitted with a corresponding image in response to the visual search query.

Images of an object may be captured via a camera at a mobile computing device at different viewpoints. The images may be used to identify components of the object and to identify damage estimates estimating damage to some or all of the components. Capture coverage levels corresponding with the components may be determined, and then recording guidance may be provided for capturing additional images to increase the capture coverage levels.

A plurality of images of an object may be processed. The plurality of images may comprise a first image having an annotation. The plurality of images may further comprise second images. The annotation may be absent from the second images. A placement of the annotation for the second images such that the annotation is configured to be included in the second images may be automatically determined. The second images may be caused to include the annotation in accordance with the determined placement. The annotated second images may be stored on a storage device.

According to various embodiments, techniques and mechanisms are provided to enhance damage detection and assessment for objects such as vehicles. In some implementations, a striped lighting tunnel is configured with cameras for capture of object or vehicle surface images when illuminated by striped lighting. The vehicle surface images may be captured from a variety of perspectives. Damage assessed is mapped to multiple views of a vehicle. In some embodiments, a damage likelihood score is generated and if the damage likelihood score reaches a threshold, a heatmap visual representation may be generated and included in one or more views to present the type, likelihood, and extent of damage.

According to various embodiments, techniques and mechanisms are provided to enhance damage detection for objects such as vehicles. In some implementations, a striped lighting tunnel is configured with cameras for capture of object or vehicle surface images when illuminated by striped lighting. The vehicle surface images may be captured from a variety of perspectives. Striped lighting capture may be implemented using gaps between uniform lighting panels and the uniform lighting panels themselves, along with high speed cameras capable of taking more than 120 frames per second. The gap width between uniform lighting panels may be determined using an inverse relationship with frame rate. Damage may be determined using striped lighting illuminated vehicle surface images to identify the type, likelihood, and extent of damage.

According to various embodiments, a variably switched striped lighting pattern panel is provided. A panel may include numerous rows of light emitting diodes (LEDs). Alternating rows are controlled separately by a microcontroller so that a first set can be switched on while a second set switched off to provide striped lighting. On the other hand, both sets can be switched on to provide for uniform lighting. Barriers may be provided between the rows of LEDs while a diffuser is placed in front of the LEDs. The variably switched striped lighting pattern panels can be used in a vehicle tunnel in a variety of configurations to allow for vehicle image capture in varying conditions for damage assessment.

A multi-view interactive digital media representation (MVIDMR) of an object can be generated from live images of an object captured from a camera. Selectable tags can be placed at locations on the object in the MVIDMR. When the selectable tags are selected, media content can be output which shows details of the object at location where the selectable tag is placed. A machine learning algorithm can be used to automatically recognize landmarks on the object in the frames of the MVIDMR and a structure from motion calculation can be used to determine 3-D positions associated with the landmarks. A 3-D skeleton associated with the object can be assembled from the 3-D positions and projected into the frames associated with the MVIDMR. The 3-D skeleton can be used to determine the selectable tag locations in the frames of the MVIDMR of the object.

A set of images may be captured by a camera as the camera moves along a path through space around an object. Then, a smoothed function (e.g., a polynomial) may be fitted to the translational and/or rotational position in space. For example, positions in a Cartesian coordinates pace may be determined for the images. The positions may then be transformed to a polar coordinate space, in which a trajectory along the points may be determined, and the trajectory transformed back into the Cartesian space. Similarly, the rotational position of the images may be smoothed, for instance by fitting a loss function. Finally, one or more images may be transformed to more closely align a viewpoint of the image with the fitted translational and/or rotational positions.

A computing system configured to process a plurality of intermediate outputs from machine learning models to generate final outputs may be maintained. A combined confidence metric that reflects a probability that the final outputs are accurate may be determined based on the intermediate outputs. Suitability of the combined outputs as inputs for a further model may be determined using the combined confidence metrics. The suitable combined outputs may be caused to be used as inputs for the further model.