An example method is described herein. The method includes executing dense feature matching on an image pair that is down sampled to obtain a first set of feature correspondences for each pixel of the down sampled image pair. The method also includes calculating a neighborhood correspondence based on the first set of feature correspondences for each pixel in a first image of the image pair. Further, the method includes executing sparse feature matching on stereoscopic patch pairs from the image pair based on the neighborhood correspondence for each pixel to obtain correspondence estimates for each stereoscopic patch pair. Finally, the method includes refining the correspondence estimates for each stereoscopic patch pair to obtain a semi-dense set of feature correspondences by applying a geometric constraint to the correspondence estimates and retaining correspondences that satisfy the geometric constraint.

A flight device and a flight control method are disclosed. The method includes: acquiring an image and a height of the flight device; determining a scene in which the flight device is currently located; calculating an image X-offset and an image Y-offset of a second image frame of two adjacent image frames relative to a first image frame of the two adjacent image frames; compensating for the image X-offset and the image Y-offset according to the acceleration and the angular velocity of the flight device to obtain image correction offsets comprising a corrected image X-offset and a corrected image Y-offset; calculating an X-offset and a Y-offset in world coordinates corresponding to the image correction offsets; and deriving a velocity of the flight device.

The present invention discloses a method for retrieving atmospheric aerosol based on statistical segmentation. Firstly a multi-band remote sensing image including an apparent reflectance and an aerosol optical thickness look-up table corresponding to a retrieval band is obtained, then pixels are partitioned and screened according to apparent reflectance segments of a mid-infrared 2.1 micrometer band. After that the retained pixel sets are further partitioned and screened according to the apparent reflectance segments of the mid-infrared 1.6 micrometer band. Finally the obtained pixel sets are partitioned into two categories according to the pixel number, one category including pixels having more pixels, the other including those with less pixels.

The category with more pixels is taken as the reference part for retrieval. Specifically, the pixel sets are first searched for the clean segment, then the ground surface reflectivity of the clean segment is taken as the ground surface reflectivity of the whole pixel set, thereby obtaining the aerosol thickness value through retrieval. After that these pixels are taken as references to perform retrieval on the other category.

The present invention can improve accuracy and resolution of the retrieval result of the bright ground surface area, and is applied to a wider range.

In implementations of multi-resolution color-based image search, an image search system determines a color vector for a query image based on a color histogram of the query image by concatenating two color histograms having different resolutions. The image search system can compute distance measures between the color vector of the query image and color vectors of candidate images. The image search system can select one or more of the candidate images to return based on the distance measures utilizing the distance measures as indication of color similarity of the candidate images to the query image.

A state recognition apparatus includes circuitry configured to: recognize a monitoring target based on a captured image; recognize a motion of each part of the recognized monitoring target; recognize a state of the monitoring target based on the recognized motion of each part; and output a state recognition result indicating the recognized state of the monitoring target.

Systems, methods, and computer program products for identifying a candidate product in an electronic marketplace based on a visual comparison between candidate product image visual content and input query image visual content. Embodiments generate and store descriptive image signatures from candidate product images or selected portions of such images. A subsequently calculated visual similarity measure serves as a visual search result score for the candidate product in comparison to an input query image. Any number of images of any number of candidate products may be analyzed, such as for items available for sale in an online marketplace. Image analysis results are stored in a database and made available for subsequent automated on-demand visual comparisons to an input query image. The embodiments enable substantially real time visual based product searching of a potentially vast catalog of items.

One of the objects of the present invention is to provide an image processing device that is able to track feature points of object accurately even under occlusion.

An image processing device according to one of the exemplary aspects of the present invention includes: feature extraction means for extracting features from an image based on a shape which includes sub parts each being one or more feature points each being correlated with a position of an object imaged in the image; alignment means for deriving an estimated shape by applying, to an initial shape, a model in models which include a global model representing a relation between the feature and the motion of the shape, and local models each representing a relation between the feature and the motion of a sub part in the sub parts; occlusion detection means for detecting, in the sub parts, an occluded sub part by comparing the estimated shape from a previous image and the estimated shape which is derived from the image by applying the global model; and estimation means for selecting, in the local models, local models each restricts the motion of the sub part that is detected as the occluded sub part, and combining estimated shape derived by applying the selected local models to generate a final output.

Techniques are disclosed for analyzing and learning behavior in an acquired stream of video frames. In one embodiment, a trajectory analyzer clusters trajectories of objects depicted in video frames and builds a trajectory model including the trajectory clusters, a prior probability of assigning a trajectory to each cluster, and an intra-cluster probability distribution indicating the probability that a trajectory mapping to each cluster is least various distances away from the cluster. Given a new trajectory, a score indicating how unusual the trajectory is may be computed based on the product of the probability of the trajectory mapping to a particular cluster and the intra-cluster probability of the trajectory being a computed distance from the cluster. The distance used to match the trajectory to the cluster and determine intra-cluster probability is computed using a parallel Needleman-Wunsch algorithm, with cells in antidiagonals of a matrix and connected sub-matrices being computed in parallel.

An information processing device according to the present invention includes: a proper identifier output unit which outputs proper identifiers for identifying learning images; a feature vector calculation unit which calculates feature vectors of at least a part of patches included in registered patches that are registered in a dictionary for compositing a restored image; and a search similarity calculation unit which calculates a similarity calculation method that classifies the proper identifiers to be given to the registered patches based on the feature vectors.

An image processing device includes a processor; and a memory which stores a plurality of instructions, which when executed by the processor, cause the processor to execute: obtaining a first image and a second image; determining a first direction of a certain first reference point of a first region of a user included in the first image and a second direction of a certain second reference point of a second region of the user included in the second image; extracting a first feature quantity from the first region and a second feature quantity from the second region; generating a third feature quantity from the first feature quantity or a fourth feature quantity from the second feature quantity on the basis of the first direction or the second direction; and authenticating the user on the basis of similarity between the first feature quantity or the third feature quantity.