Abstract: 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.

Abstract: Disclosed systems and methods generate page segmented documents from unstructured vector graphics documents. The page segmentation application executing on a computing device receives as input an unstructured vector graphics document. The application generates an element proposal for each of many areas on a page of the input document tentatively identified as being page elements. The page segmentation application classifies each of the element proposals into one of a plurality of defined type of categories of page elements. The page segmentation application may further refine at least one of the element proposals and select a final element proposal for each element within the unstructured vector document. One or more of the page segmentation steps may be performed using a neural network.

Abstract: Methods and systems are provided for generating mattes for input images. A neural network system can be trained where the training includes training a first neural network that generates mattes for input images where the input images are synthetic composite images. Such a neural network system can further be trained where the training includes training a second neural network that generates refined mattes from the mattes produced by the first neural network. Such a trained neural network system can be used to input an image and trimap pair for which the trained system will output a matte. Such a matte can be used to extract an object from the input image. Upon extracting the object, a user can manipulate the object, for example, to composite the object onto a new background.

Abstract: Disclosed systems and methods generate page segmented documents from unstructured vector graphics documents. The page segmentation application executing on a computing device receives as input an unstructured vector graphics document comprising drawing commands. The application generates an element proposal for each of many areas on a page of the input document tentatively identified as being page elements. Each of the element proposals may be generated at least in part based on the drawing commands. The page segmentation application classifies each of the element proposals into one of a plurality of defined type of categories of page elements at least in part based on the drawing commands. The page segmentation application may further refine at least one of the element proposals and select a final element proposal for each element within the unstructured vector document. One or more of the page segmentation steps may be performed using a neural network.

Abstract: Methods and systems are provided for generating mattes for input images. A neural network system can be trained where the training includes training a first neural network that generates mattes for input images where the input images are synthetic composite images. Such a neural network system can further be trained where the training includes training a second neural network that generates refined mattes from the mattes produced by the first neural network. Such a trained neural network system can be used to input an image and trimap pair for which the trained system will output a matte. Such a matte can be used to extract an object from the input image. Upon extracting the object, a user can manipulate the object, for example, to composite the object onto a new background.

Abstract: Techniques are disclosed herein that enable digital images to be segmented based on a user's semantic input. In other words, given an input image of a person walking a dog adjacent to a tree, a user can simply provide the semantic input “dog” and the system will segment the dog from the other elements in the image. If the user provides other semantic input, such as “person” or “tree”, the system will segment the person or the tree, respectively, from the same image. Using semantic input advantageously eliminates any need for a user to directly interact with the input image through a tedious process of painting brush strokes, tracing boundaries, clicking target points, and/or drawing bounding boxes. Thus semantic input represents an easier and more intuitive way for users to interact with an image segmentation interface, thereby enabling novice users to take advantage of advanced image segmentation techniques.

Abstract: Techniques are disclosed herein that enable digital images to be segmented based on a user's semantic input. In other words, given an input image of a person walking a dog adjacent to a tree, a user can simply provide the semantic input “dog” and the system will segment the dog from the other elements in the image. If the user provides other semantic input, such as “person” or “tree”, the system will segment the person or the tree, respectively, from the same image. Using semantic input advantageously eliminates any need for a user to directly interact with the input image through a tedious process of painting brush strokes, tracing boundaries, clicking target points, and/or drawing bounding boxes. Thus semantic input represents an easier and more intuitive way for users to interact with an image segmentation interface, thereby enabling novice users to take advantage of advanced image segmentation techniques.

Abstract: Techniques are disclosed herein that enable digital images to be segmented based on a user's semantic input. In other words, given an input image of a person walking a dog adjacent to a tree, a user can simply provide the semantic input “dog” and the system will segment the dog from the other elements in the image. If the user provides other semantic input, such as “person” or “tree”, the system will segment the person or the tree, respectively, from the same image. Using semantic input advantageously eliminates any need for a user to directly interact with the input image through a tedious process of painting brush strokes, tracing boundaries, clicking target points, and/or drawing bounding boxes. Thus semantic input represents an easier and more intuitive way for users to interact with an image segmentation interface, thereby enabling novice users to take advantage of advanced image segmentation techniques.

Abstract: Techniques are disclosed herein that enable digital images to be segmented based on a user's semantic input. In other words, given an input image of a person walking a dog adjacent to a tree, a user can simply provide the semantic input “dog” and the system will segment the dog from the other elements in the image. If the user provides other semantic input, such as “person” or “tree”, the system will segment the person or the tree, respectively, from the same image. Using semantic input advantageously eliminates any need for a user to directly interact with the input image through a tedious process of painting brush strokes, tracing boundaries, clicking target points, and/or drawing bounding boxes. Thus semantic input represents an easier and more intuitive way for users to interact with an image segmentation interface, thereby enabling novice users to take advantage of advanced image segmentation techniques.

Abstract: A process for forming a protective polyimide layer over a semiconductor substrate includes the steps of curing a deposited polyamic acid layer at a temperature which is sufficient to reduce the etch rate of the acid layer when subsequently exposed to a developer. After formation of a photoresist masking layer over the polyamic acid, the substrate is exposed to a developer to define a plurality of bonding pad openings therein. The developer permeates into the acid layer to form a salt in the regions beneath the openings. Subsequent hardbaking imidizes the polyamic acid, but not the salt regions. Removing the photoresist layer also develops the polyimide which removes the salt regions to expose the underlying bonding pads.

Abstract: A failure detector detects breakage in a machine tool cutting element, such as a drill bit, by use of a accelerometer supported against a workpiece being operated on by the cutting element. Vibrations are induced in the workpiece by the cutting operation, the vibrations having spectral components in the audio frequency range which are characteristic of proper operation of the cutting element, the spectral components changing upon a breakage of the cutting element. A voltage produced by a positioning mechanism of the cutting element identifies the location of the cutting element relative to the workpiece. A bandpass filter separates the spectral components associated with proper operation from spectral components associated with faulty operation. Logic circuitry signals the joint occurrence of the spectral components of proper operation with the position of the cutting element at the site of the workpiece so as to permit the signaling of a cutting tool failure during operation on the workpiece.