Nippon Telegraph and Telephone Corporation

Atsushi Sagata

Shingo Ando

Yasuhiro Yao

Hitoshi Niigaki

Kana Kurata

Atsushi Shimizu

Hiroshi Watanabe

Kazuto Kamikura

Jun Shimamura

Kazuhiko Murasaki

Hirohisa Jozawa

Yukito Watanabe

Hiroyuki Oshida

Seishi Takamura

Tomoya Shimizu

Masaaki Inoue

Yukihiro Goto

Ryuji Honda

Takashi Hosono

Mitsuo Ikeda

Shigehiro Matsuda

Masaki Waki

Jirou Naganuma

Takayuki Umeda

Kaori Kumagai

Shunsuke Tsukatani

Hiroaki Matsuda

A point group including a small number of points that have been assigned labels is taken as an input to assign labels to points that have not been assigned labels. In a label estimation apparatus for estimating a label to be assigned to a point that has not been labeled using a label of a point that has been labeled among points included in a point group, a confidence derivation unittakes a point that has not been labeled within a point group including a point that has been labeled and the point that has not been labeled as a target point and estimates a class of the target point and a likelihood indicating a confidence of an estimation result of the class from a set of points included in the point group, a priority derivation unitobtains a distance between the target point and a point that has been assigned the same label as a label corresponding to the estimated class as a priority used to determine whether the estimated class is appropriate, and a label determination unitdetermines whether the estimated class is appropriate using at least an index based on the distance.

Label estimation device, label estimation method, and label estimation program

Annotation can be easily performed on a three-dimensional point cloud and a working time can be reduced. An interface unitdisplays a point cloud indicating a three-dimensional point on an object, and receives designation of a three-dimensional point indicating an annotation target object and designation of a three-dimensional point not indicating the annotation target object. A candidate cluster calculation unitcalculates a value of a predetermined evaluation function indicating a likelihood of a point cloud cluster being the annotation target object based on the designation of a three-dimensional point for point cloud clusters obtained by clustering the point clouds. A cluster selection and storage designation unitcauses the interface unitto display the point cloud clusters in descending order of the value of the evaluation function, and receives a selection of a point cloud cluster to be annotated. An annotation execution unitexecutes annotation indicating the annotation target object for each three-dimensional point included in the selected point cloud cluster.

Point cloud annotation device, method, and program

An identification result explanation device calculates, for each of feature maps output individually from a plurality of filters used in specified layers of a CNN, a weight representing a degree of association with a result of identification for an input image by the CNN. The identification result explanation device outputs transposed feature maps obtained by transposing, for each of clusters, the feature maps in the cluster based on a result of classification of each of the feature maps for the input image classified into any of the clusters and on the weight calculated for each of the feature maps. The identification result explanation device uses the transposed feature maps in each of the clusters to retrieve, in a storage unit, each of the clusters including the feature maps linked to the same filters as those linked to the transposed feature maps for the input image and selects.

Image reaction map generation device, image reaction map generation method, and image reaction map generation program

Provided is a point cloud analysis device that curbs a decrease in model estimation accuracy due to a laser measurement point cloud. A clustering unit () clusters a point cloud representing a three-dimensional point on an object obtained by a measurement unit mounted on a moving body and performing measurement while scanning a measurement position, within a scan line, to obtain a point cloud cluster. A central axis direction estimation unit () estimates a central axis direction based on the point cloud cluster. A direction-dependent local effective length estimation unit () estimates a local effective length based on an estimated central axis direction and an interval of scan lines, the local effective length being a length when a length of projection of the point cloud cluster in a central axis direction for each of the point cloud clusters is interpolated by an amount of a loss part of the point cloud.

Point cloud analysis device, method, and program

A three-dimensional point cloud label learning and estimation device includes: a clustering unit that clusters a three-dimensional point cloud into clusters; a learning unit that makes a neural network learn to estimate a label corresponding to an object to which points contained in each of the clusters belong; and an estimation unit that estimates a label for the cluster using the neural network learned at the learning unit. In the three-dimensional point cloud label learning and estimation device, the neural network uses a total sum of sigmoid function values (sum of sigmoid) when performing feature extraction on the cluster.

Three-dimensional point cloud label learning estimation device, three-dimensional point cloud label learning estimation method, and 3D point cloud label learning estimation program

Labels can be accurately identified even for an image with a resolution not used in training data. Based on an input image, a resolution of the input image, and a resolution of a training image used for training a trained model of assigning labels to pixels of an image, a plurality of low-resolution images are generated from the input image by using a plurality of shift amounts for a pixel correspondence between the input image and the respective low-resolution images with a resolution corresponding to the training image, the low-resolution images are input to the trained model, a plurality of low-resolution label images is output in which pixels of the respective low-resolution images are assigned labels, and a label image is output in which labels for pixels of the input image are obtained, based on the shift amounts used for generating the low-resolution images and the low-resolution label images.

Image processing device, method, and program

An object is to make it possible to train an image recognizer by efficiently using training data that does not include label information. A determination unitcauses repeated execution of the followings. A feature representation model for extracting feature vectors of pixels is trained such that an objective function is minimized, the objective function being expressed as a function that includes a value that is based on a difference between a distance between feature vectors of pixels labeled with a positive example label and a distance between a feature vector of a pixel labeled with the positive example label and a feature vector of an unlabeled pixel, and a value that is based on a difference between a distance between a feature vector of a pixel labeled with the positive example label and a feature vector of an unlabeled pixel and a distance between a feature vector of a pixel labeled with the positive example label and a feature vector of a pixel labeled with a negative example label, and based on a distribution of feature vectors corresponding to the positive example label, a predetermined number of labels are given based on the likelihood that each unlabeled pixel is a positive example.

Image classifier learning device, image classifier learning method, and program

An object included in a low-resolution image can be recognized with a high degree of precision. An acquisition unit acquires, from a query image, an increased-resolution image, which is acquired by increasing the resolution of the query image, by performing pre-learned acquisition processing for increasing the resolution of an image. A feature extraction unit, using the increased-resolution image as input, extracts a feature vector of the increased-resolution image by performing pre-learned extraction processing for extracting a feature vector of an image. A recognition unit recognizes an object captured on the increased-resolution image on the basis of the feature vector of the increased-resolution image and outputs the recognized object as the object captured on the query image.

Object recognition device, object recognition learning device, method, and program

It is possible to estimate a slack level accurately in consideration of a shape of a deformed cable. A point cloud analysis device sets a plurality of regions of interest obtained by window-searching a wire model including a quadratic curve model representing a cable obtained from a point cloud consisting of three-dimensional points on an object, the region of interest being divided into a first region and a second region. The point cloud analysis device compares information on the first region with information on the second region based on the point cloud included in the region of interest and the quadratic curve model for each of the plurality of regions of interest, calculates a degree of division boundary representing a degree to which a division position between the first region and the second region of the plurality of regions of interest is a branch point of the cable, and detects a division boundary point that is a branch point of a cable represented by the quadratic curve model based on the degree of division boundary calculated for each of the plurality of regions of interest.

Point cloud analysis device, estimation device, point cloud analysis method, and program

Dense depth information can be generated using only a monocular image and sparse depth information. A depth hyper-resolving apparatusincludes: an input data processing unitthat outputs a hierarchical input image and hierarchical input depth information by resolution conversion in accordance with a predetermined number of tiers for an input image and input depth information; a depth continuity estimation unitthat derives hierarchical estimated depth continuity based on the hierarchical input image; a depth continuity mask deriving unitthat outputs a hierarchical depth continuity mask representing values of locations depending on whether a depth is continuous based on the hierarchical input image and the hierarchical estimated depth continuity; and a cost function minimization unitthat derives hyper-resolved depth information to minimize a cost function expressed by using the hierarchical input depth information, the hierarchical depth continuity mask, and the hyper-resolved depth information.

Depth superresolution device, depth superresolution method, and program

A mesh structure facility detection device detects data corresponding to a mesh structure facility from three-dimensional structure data representing a space including an outer shape of an object, and projects the three-dimensional structure data in a predetermined direction to obtain two-dimensional structure data; and detects a point included in a region in which the two-dimensional structure data has a density of more than or equal to a predetermined threshold value as a point corresponding to the mesh structure facility.

Mesh structure equipment detection apparatus, mesh structure equipment detection method and program

A list for accurately identifying objects with different sizes that are the same attributes can be generated automatically. A classification unit classifies, from a group of images consisting of images including objects with any of a plurality of attributes, each of the images including the objects that have an identical attribute and have different real sizes into an identical cluster. An output unit outputs each of clusters into which at least two of the images are classified as a list of images with different real sizes of the objects for the identical attribute.

List generation device, photographic subject identification device, list generation method, and program

A retrieval apparatus includes a first retrieval unit, a second retrieval unit, and an integration unit that calculates an integrated similarity by integrating a first similarity calculated by the first retrieval unit and a second similarity calculated by the second retrieval unit. For similarities between a basic image, an image similar to the basic image, and an image dissimilar to the basic image of the reference images, at least the feature extraction of the first retrieval unit is learned such that a margin based on a second similarity between the basic image and the similar image and a second similarity between the basic image and the dissimilar image increases as the second similarity between the basic image and the dissimilar image increases relative to the second similarity between the basic image and the similar image.

Learning device, search device, learning method, search method, learning program, and search program

A motion vector predictive encoding method, a motion vector decoding method, a predictive encoding apparatus, a decoding apparatuses, and storage media storing motion vector predictive encoding and decoding programs are provided, thereby reducing the amount of generated code with respect to the motion vector, and improving the efficiency of the motion-vector prediction. If the motion-compensating mode of the target small block to be encoded is the global motion compensation, the encoding mode of an already-encoded small block is the interframe coding mode, and the motion-compensating mode of the already-encoded small block is the global motion compensation, then the motion vector of the translational motion model is determined for each pixel of the already-encoded small block, based on the global motion vector (steps S-S). Next, the representative motion vector is calculated as the predicted vector, based on the motion vector of each pixel of the already-encoded small block (step S). Finally, the prediction error is calculated for each component of the motion vector and each prediction error is encoded (steps Sand S).

Motion vector predictive encoding and decoding method using prediction of motion vector of target block based on representative motion vector

A video signal generation apparatus for decoding a bit stream to generate a video signal, includes: a writing unit configured to write a flag value indicating whether a decoding target block has been coded by intra-coding or inter-coding in a specified bit position on one bit plane having a size the same as that of the decoding target block, to write information indicating a coding division form of the decoding target block in a specified common area on the one bit plane, and to write the coding information in a specified switch interpretation area on the one bit plane according to a data format conforming to the intra-coding if the decoding target block has been coded by intra-coding, and to write the coding information in the specified switch interpretation area according to a data format conforming to the inter-coding if the decoding target block has been coded by inter-coding; and a generation unit configured to generate a video signal by superimposing the one bit plane on a decoded image of the decoding target block.

Video signal generation apparatus, video signal generation method, video signal generation program, and computer readable recording medium recording the program

Growing community of inventors

Tokyo, Japan

Atsushi Sagata