Disclosed is an apparatus and method for calculating a raw-cost necessary for combining images into one image by matching of stereo images. The raw-cost calculation apparatus includes an image acquirer, a window generator, a window mask generator, a window masker, and a raw-cost calculator. In the raw-cost calculation apparatus and method, a raw cost may be calculated by using an adaptive window mask so that accurate 3D information may be obtained on the boundary of thin structures even when stereo images are matched and combined.

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 system for image processing that matches a model image with an input image. The matching process includes using a feature location index for the model image.

The invention provides an image processing device, an image processing method, and an image processing program capable of correctly discriminating a type of a test object, even when similar model images have been registered. The image processing device includes: a hardware that holds a feature amount obtained from model images of a plurality of reference objects belonging to mutually different types; a region determination module that determines a non-common region as a region indicating a feature amount different from those of other objects, within a model image of each object; and a discrimination module that discriminates which type an object included in an input image belongs to, by using a feature amount corresponding to a non-common region of the object out of feature amounts of objects.

A hardness tester includes a memory associating and storing a parts program having defined measurement conditions with respect to a sample, including a test position, and an image file acquired by capturing an image of the shape of the sample; an image acquirer acquiring image data of the sample to be measured; a pattern matcher performing a pattern matching process on the image data of the sample using the image file associated with the parts program; a determiner determining whether an image file exists which has a shape related to the image data of the sample; a retriever retrieving the parts program associated with the image file having a related shape; and a measurer measuring hardness of the sample based on the retrieved parts program.

Object detection techniques are described. In one or more implementations, a plurality of images are received by a computing device. The plurality of images are analyzed by the computing device to detect whether the images include, respectively, a depiction of an object. If an object is found in a first image, the locations, angles and scales for object detection can be further restricted in a second one. If an object is not found in a first one of the image, different portions of a second one of the images are analyzed for object detection.

In implementations of the subject matter as described herein, there is provided a method for forgery detection of a face image. Subsequent to inputting a face image, it is detected whether a blending boundary due to the blend of different images exists in the face image, and then a corresponding grayscale image is generated based on a result of the detection, where the generated grayscale image can reveal whether the input face image is formed by blending different images. If a visible boundary corresponding to the blending boundary exists in the generated grayscale image, it indicates that the face image is a forged image; on the contrary, if the visible boundary does not exist in the generated grayscale image, it indicates that the face image is a real image.

A method of labelling a water surface within an image is provided. The method includes: receiving image data of the image generated by a camera, the image including at least one water surface; receiving water surface extension data, the water surface extension data being representative of an area over which the water surface extends in the real world; matching the water surface extension data to the image data based on a spatial relationship between the camera and the area of the water surface; and labelling the water surface in the image based on the matched water surface extension data.

An image processing apparatus includes at least one processor that is configured to: acquire projection images obtained by imaging a subject with radiation at a plurality of angles; reconstruct a plurality of tomographic images from the acquired projection images; detect a structure of interest from the plurality of reconstructed tomographic images; derive a region of the projection image corresponding to the detected structure of interest; and determine, by using the derived region of the projection image, whether or not the tomographic image from which the structure of interest is detected is a tomographic image corresponding to a focal plane of the structure of interest.

Systems, devices, and methods are described herein for improving object detection frameworks. Proposed regions can be used to identify similar images from a novel image set. Once identified, a weighted average of the feature representations of the similar images and/or a probability distribution of the classification labels for those images can be generated. The weighted average of the feature representations and/or the probability distribution can be used to steer the predicted classification confidence and/or predicted bounding box coordinates of the object detection framework. The disclosed techniques can be easily integrated with the object detect framework to improve the accuracy of its predictions without adding additional trainable parameters so as to refrain from adding complexity to the learning process.