A method includes obtaining, from an image sensor, image data of a real-world scene; obtaining, from a depth sensor, sparse depth data of the real-world scene; and passing the image data to a first neural network to obtain one or more object regions of interest (ROIs) and one or more feature map ROIs. Each object ROI includes at least one detected object. The method also includes passing the image data and sparse depth data to a second neural network to obtain one or more dense depth map ROIs; aligning the one or more object ROIs, one or more feature map ROIs, and one or more dense depth map ROIs; and passing the aligned ROIs to a fully convolutional network to obtain a segmentation of the real-world scene. The segmentation contains one or more pixelwise predictions of one or more objects in the real-world scene.

Described is a system and method for enabling dynamic selection of a search input. For example, rather than having a static search input box, the search input may be dynamically positioned such that it encompasses a portion of displayed information. An image segment that includes a representation of the encompassed portion of the displayed information is generated and processed to determine an object represented in the portion of the displayed information. Additional images with visually similar representations of objects are then determined and presented to the user.

A device include on-board memory, an applications processor, a digital signal processor (DSP) cluster, a configurable accelerator framework (CAF), and at least one communication bus architecture. The communication bus communicatively couples the applications processor, the DSP cluster, and the CAF to the on-board memory. The CAF includes a reconfigurable stream switch and data volume sculpting circuitry, which has an input and an output coupled to the reconfigurable stream switch. The data volume sculpting circuitry receives a series of frames, each frame formed as a two dimensional (2D) data structure, and determines a first dimension and a second dimension of each frame of the series of frames. Based on the first and second dimensions, the data volume sculpting circuitry determines for each frame a position and a size of a region-of-interest to be extracted from the respective frame, and extracts from each frame, data in the frame that is within the region-of-interest.

A method, apparatus and system for zero shot object detection includes, in a semantic embedding space having embedded object class labels, training the space by embedding extracted features of bounding boxes and object class labels of labeled bounding boxes of known object classes into the space, determining regions in an image having unknown object classes on which to perform object detection as proposed bounding boxes, extracting features of the proposed bounding boxes, projecting the extracted features of the proposed bounding boxes into the space, computing a similarity measure between the projected features of the proposed bounding boxes and the embedded, extracted features of the bounding boxes of the known object classes in the space, and predicting an object class label for proposed bounding boxes by determining a nearest embedded object class label to the projected features of the proposed bounding boxes in the space based on the similarity measures.

An image processing apparatus according to the present invention includes: an extracting unit configured to extract a candidate image, which is an image of a candidate region specified in accordance with a preset criterion, from a target image to be a target for an annotation process, and also extract a corresponding image, which is an image of a corresponding region corresponding to the candidate region, from a reference image that is an image corresponding to the target image; a displaying unit configured to display the candidate image and the corresponding image so as to be able to compare the images with each other; and an input accepting unit configured to accept input of input information for the annotation process for the candidate image.

An image processing method includes performing additional image feature extraction on a training source face image to obtain a source additional image feature, performing identity feature extraction on the training source face image to obtain a source identity feature, inputting a training template face image into an encoder in a to-be-trained face swapping model to obtain a face attribute feature, inputting the source additional image feature, the source identity feature, and the face attribute feature into a decoder in the face swapping model for decoding to obtain a decoded face image, obtaining a target model loss value based on an additional image difference between the decoded face image and a comparative face image, and adjusting the model parameters of the encoder and the decoder based on the target model loss value to obtain the trained face swapping model.

There is provided a system and method for mask inspection, comprising: obtaining a plurality of images, each representative of a respective part of the mask; generating a CD map of the mask comprising a plurality of composite values of a CD measurement of a POI respectively derived from the plurality of images, comprising, for each given image: dividing the given image into a plurality of sections; searching for the POI in the plurality of sections, giving rise to a set of sections, each with presence of at least one of the POI therein; for each section, obtaining a value of the CD measurement using a printing threshold, giving rise to a set of values of the CD measurement corresponding to the set of sections; and combining the set of values to a composite value of the CD measurement corresponding to the given image.

Provided are a depth image processing method, and a small obstacle detection method and system. The method comprises calibration of sensors, distortion and epipolar rectification, data alignment, and sparse stereo matching. The depth image processing method and the small obstacle detection method and system of the present invention only requires execution of sparse stereo matching on hole portions of a structured light depth image, and do not requires stereo matching of the entire image, thereby significantly reducing the overall computation load for processing a depth image, and enhancing the robustness of a system.

A method is disclosed that includes receiving, by a processing device, a plurality of images of a test assembly. The processing device selects a component in the test assembly and an image of the plurality of images of the test assembly as received. For the component as selected and the image as selected, the processing device compares a plurality of portions of the image as selected to a corresponding plurality of portions of a corresponding profile image and computing a matching score for each of the plurality of portions. The processing device selects a largest matching score from the matching score for each of the plurality of portions as a first matching score for the component as selected and the image as selected. The first matching score is stored for the component as selected and the image as selected.

When there is an area that cannot be confirmed as a white line based on an image acquired by a camera, it is determined that the area that cannot be confirmed as the white line is an area that can be regarded as the white line, on condition that the area that cannot be confirmed as the white line is located in an area between a first straight line connecting edges of one ends of the area that can be confirmed as the white line in a longitudinal direction of the white line and extension lines of the first straight line, and a second straight line connecting edges of the other ends and extension lines of the second straight line. For the area that can be regarded as the white line, a shape of the area in the image is set to a shape as a white line.