The present disclosure relates to systems, non-transitory computer-readable media, and methods for training and utilizing scale-diverse segmentation neural networks to analyze digital images at different scales and identify different target objects portrayed in the digital images. For example, in one or more embodiments, the disclosed systems analyze a digital image and corresponding user indicators (e.g., foreground indicators, background indicators, edge indicators, boundary region indicators, and/or voice indicators) at different scales utilizing a scale-diverse segmentation neural network. In particular, the disclosed systems can utilize the scale-diverse segmentation neural network to generate a plurality of semantically meaningful object segmentation outputs. Furthermore, the disclosed systems can provide the plurality of object segmentation outputs for display and selection to improve the efficiency and accuracy of identifying target objects and modifying the digital image.

Methods and devices that identify a silkscreen data file associated with a physical printed circuit board and use an image of the physical printed circuit board to display a virtual silkscreen over the image of the physical printed circuit board.

Methods and systems including computer programs encoded on a computer storage medium, for generating and displaying object recognition state indicators during object recognition processing of an image. In one aspect, a method includes performing object recognition on an image displayed in an application environment of an application on a user device using an object recognition model having multiple object recognition states including an identification state, where a candidate object in the image is positively identified, and one or more precursor states to the identification state, and where each of the precursor states has a different respective indicator for display within the image during the respective precursor state that visually emphasizes the candidate object and the identification state has a different respective indicator for display within the image during the identification state that visually emphasizes the positively identified object as being positively identified.

Systems and techniques are described herein for directing a user conversation to obtain an editing query, and removing and replacing objects in an image based on the editing query. Pixels corresponding to an object in the image indicated by the editing query are ascertained. The editing query is processed to determine whether it includes a remove request or a replace request. A search query is constructed to obtain images, such as from a database of stock images, including fill material or replacement material to fulfill the remove request or replace request, respectively. Composite images are generated from the fill material or the replacement material and the image to be edited. Composite images are harmonized to remove editing artifacts and make the images look natural. A user interface exposes images, and the user interface accepts multi-modal user input during the directed user conversation.

Described are techniques for determining one or more resources usable to produce an image from content data associated with existing images selected by a user. User preference data indicative of a set of images may be generated, and from the image data and resource data associated with the set of images, one or more resources used to produce the images may be determined. Output data indicative of the resources may be generated for provision to a user.

Various embodiments of the present invention relate generally to systems and methods for analyzing and manipulating images and video. According to particular embodiments, the spatial relationship between multiple images and video is analyzed together with location information data, for purposes of creating a representation referred to herein as a surround view for presentation on a device. A visual guide can provided for capturing the multiple images used in the surround view. The visual guide can be a synthetic object that is rendered in real-time into the images output to a display of an image capture device. The visual guide can help user keep the image capture device moving along a desired trajectory.

Various embodiments of the present invention relate generally to systems and methods for analyzing and manipulating images and video. According to particular embodiments, the spatial relationship between multiple images and video is analyzed together with location information data, for purposes of creating a representation referred to herein as a surround view for presentation on a device. A visual guide can provided for capturing the multiple images used in the surround view. The visual guide can be a synthetic object that is rendered in real-time into the images output to a display of an image capture device. The visual guide can help user keep the image capture device moving along a desired trajectory.

The present disclosure discloses a search method and apparatus. A specific implementation of the method comprises: receiving first information for searching, the first information for searching comprising an image for searching; determining an image attribute tag selected by a selection operation as being second information for searching in response to detecting the selection operation of the preset image attribute tag, the image attribute tag being used for indicating an attribute of the image; sending the received first information for searching and the determined second information for searching to a server, so that the server conducts a search according to the first information for searching and the second information for searching to obtain and send a first image search result; and receiving and displaying the first image search result sent by the server. This implementation improves the search efficiency.

Systems and methods are provided for implementing a computer vision model execution service. A computer vision model execution service may maintain a library of machine learning models, and may provide a standard interface for accessing a model or models. Models may implement schemas that specify an input vector and an output vector, and the computer vision model execution service may obtain or determine workflows that process input vectors through multiple computer vision models. The service further provides an interface for adding, removing, or updating models, and may provide feedback to modelers regarding the usage and performance of various machine learning models.

Methods and systems including computer programs encoded on a computer storage medium, for generating and displaying object recognition state indicators during object recognition processing of an image. In one aspect, a method includes performing object recognition on an image displayed in an application environment of an application on a user device using an object recognition model having multiple object recognition states including an identification state, where a candidate object in the image is positively identified, and one or more precursor states to the identification state, and where each of the precursor states has a different respective indicator for display within the image during the respective precursor state that visually emphasizes the candidate object and the identification state has a different respective indicator for display within the image during the identification state that visually emphasizes the positively identified object as being positively identified.