Adobe, Inc.

Peter Merrill

Paris Smaragdis

Gautham J Mysore

Minje Kim

Gregg Darryl Wilensky

Andrea L Mangini

Robert William Teele

Yifei Fan

Jiawen Chen

Ethan Tseng

Lars Jebe

Digital content stylization techniques are described that leverage a neural photofinisher to generate stylized digital images. In one example, the neural photofinisher is implemented as part of a stylization system to train a neural network to perform digital image style transfer operations using reference digital content as training data. The training includes calculating a style loss term that identifies a particular visual style of the reference digital content. Once trained, the stylization system receives a digital image and generates a feature map of a scene depicted by the digital image. Based on the feature map as well as the style loss, the stylization system determines visual parameter values to apply to the digital image to incorporate a visual appearance of the particular visual style. The stylization system generates the stylized digital image by applying the visual parameter values to the digital image automatically and without user intervention.

Neural photofinisher digital content stylization

Methods and systems are provided for identifying parameter image adjustments. In embodiments, a set of candidate parameter values associated with a parameter to be analyzed in association with an image is identified. Subsequently, the image is rendered in accordance with each candidate parameter value to generate a set of rendered images. A neural network can then be used to identify a parameter image adjustment to apply to the image based on features associated with the set of rendered images. The neural network can be trained based on a comparison of the identified parameter image adjustment and a reference parameter value associated with the parameter being analyzed.

Identifying parameter image adjustments using image variation and sequential processing

An audio and video synchronizing perceptual model is described that is based on how a person perceives audio and/or video (e.g., how the brain processes sound and/or visual content). The relative emotional impact associated with different audio portions may be employed to determine transition points to facilitate automatic synchronization of audio data to video data to create a production that achieves a particular overall emotional effect on the listener/viewer. Various processing techniques of the perceptual model may utilize perceptual characteristics within the audio portions to determine a transition point for automatic synchronization with video data.

Audio and video synchronizing perceptual model

Systems and methods disclosed create one or more digital media presentations based on impact values. In particular, in one or more embodiments, systems and methods generate impact values based on a change in a measure of energy associated with digital audio content over time. For instance, systems and methods generate impact values by calculating a difference in a measure of energy over time in relation to the amount of energy at a particular time period. Based on the generated impact values, systems and methods identify transition points in the digital audio content. Specifically, systems and methods utilize a decaying masking threshold to identify transition points from generated impact values. Moreover, systems and method utilize identified transition points to modify digital visual content displayed in conjunction with the digital audio content.

Utilizing audio digital impact to create digital media presentations

Embodiments of the present invention relate to detecting irregularities in audio, such as music. An input signal corresponding to an audio stream is received. The input signal is transformed from a time domain into a frequency domain to generate a plurality of frames that each comprises frequency information for a portion of the input signal. An irregular event in a portion of the input signal corresponding to a set of frames in the plurality of frames is identified based on a comparison of frequency information of the set of frames to the frequency information of other sets of frames of the plurality of frames. This allows an indication of the irregular event to be provided, or for the input signal to be automatically synchronized to a multimedia event.

Irregularity detection in music

Techniques are disclosed for automatic detection of dense ornamentation in music. Input data representing a piece of digitally encoded music in a time domain is converted into a spectrogram representing time-frequency coefficients in a frequency domain. The spectrogram includes column vectors of the time-frequency coefficients that correspond to time periods spanning different portions of the piece of music. A one-dimensional onset detection array is calculated based on a subset of the column vectors. Using the spectrogram and the onset detection array, a two-dimensional self-similarity matrix (SSM) is calculated based on pair-wise comparisons of elements in the onset detection array. As a result, an irregular pattern score representing the presence of dense ornamentation in the piece of music can be calculated based on a magnitude difference between a beat pattern in the music and each column of the slim SSM.

Automatic detection of dense ornamentation in music

One embodiment of the present invention provides a system that generates queries to search for matching forms in a set of forms, without requiring a pre-configured forms management system. During operation, the system obtains a form template, wherein the form template includes a form schema which specifies field names and data types for fields in form instances associated with the form template. Next, the system allows a user to specify search criteria for one or more fields defined by the form schema. Then the system generates a query based on search criteria entered by the user, and runs the query against the set of forms to identify matching forms. Finally, the system generates a list of the matching forms.

Method and apparatus for generating a query to search for matching forms

A method renders a graphical object for display on a destination raster of pixels having a destination resolution by rendering the graphical object at a resolution higher than the destination resolution to generate one or more pieces, the pieces in the aggregate representing the graphical object at the higher resolution, and each of the pieces having a position with respect to the destination raster, whereby each piece overlaps one or more pixels; for each of the pieces, calculating a contribution made by the piece to each of one or more pixels according to a selected algorithm, the algorithm being selected by a process applying a plurality of factors; and aggregating all the contributions made to each pixel to determine a value for the pixel.

Growing community of inventors

Sunnyvale, CA, United States of America

Peter Merrill