Abstract: A back-propagation significance detection method based on depth map mining, comprising: for an input image Io, at a preprocessing phase, obtaining a depth image Id and an image Cb with four background corners removed of the image Io; at a first processing phase, carrying out positioning detection on a significant region of the image by means of the obtained image Cb with four background corners removed and the obtained depth image Id to obtain the preliminary detection result S1 of a significant object in the image; then carrying out depth mining on a plurality of processing phases of the depth image Id to obtain corresponding significance detection results; and then optimizing the significance detection result mined in each processing phase by means of a back-propagation mechanism to obtain a final significance detection result map. The method can improve the detection accuracy of the significance object.

Abstract: Provided is an a priori constraint and outlier suppression based image deblurring method. A convolution model is used for fitting a blurring process of a clear image and the blurred image I is restored, so that the purpose of image deblurring is achieved. The method comprises an evaluation process of the significant structure of a blurred image, a process of blurring kernel estimation and outlier suppression, and a process of restoring the blurred image by non-blind deconvolution. A structure in the blurred image is obtained by use of L0 norm constraint and heavy-tailed a priori information. The L0 norm constraint is used to evaluate the blurring kernel. The evaluated blurring kernel is subjected to outlier suppression. The final restored image is obtained by using a non-blind deconvolution algorithm. The present invention can prominently improve the restoration level of the blurred image.

Abstract: An intra-frame prediction-based point cloud attribute compression method. A new block structure-based intra-frame prediction scheme is provided for point cloud attribute information, where four prediction modes are provided to reduce information redundancy among different coding blocks as much as possible and improve point cloud attribute compression performance. The method comprises: performing point cloud attribute color space conversion; dividing a point cloud by using a K-dimensional (KD) tree to obtain coding blocks; performing block structure-based intra-frame prediction; performing intra-frame prediction mode division; performing conversion, uniform quantization, and entropy coding.

Abstract: A bidirectional image-text retrieval method based on a multi-view joint embedding space includes: performing retrieval with reference to a semantic association relationship at a global level and a local level, obtaining the semantic association relationship at the global level and the local level in a frame-sentence view and a region-phrase view, and obtaining semantic association information in a global level subspace of frame and sentence in the frame-sentence view, obtaining semantic association information in a local level subspace of region and phrase in the region-phrase view, processing data by a dual-branch neural network in the two views to obtain an isomorphic feature and embedding the same in a common space, and using a constraint condition to reserve an original semantic relationship of the data during training, and merging the two semantic association relationships using multi-view merging and sorting to obtain a more accurate semantic similarity between data.

Abstract: Disclosed is a deep discriminative network for person re-identification in an image or a video. Concatenation are carried out on different input images on a color channel by constructing a deep discriminative network, and an obtained splicing result is defined as an original difference space of different images. The original difference space is sent into a convolutional network. The network outputs the similarity between two input images by learning difference information in the original difference space, thereby realizing person re-identification. The features of an individual image are not learnt, and concatenation are carried out on input images on a color channel at the beginning, and difference information is learnt on an original space of the images by using a designed network. By introducing an Inception module and embedding the same into a model, the learning ability of a network can be improved, and a better differentiation effect can be achieved.

Abstract: The present application discloses a cross-media retrieval method based on deep semantic space, which includes a feature generation stage and a semantic space learning stage. In the feature generation stage, a CNN visual feature vector and an LSTM language description vector of an image are generated by simulating a perception process of a person for the image; and topic information about a text is explored by using an LDA topic model, thus extracting an LDA text topic vector. In the semantic space learning phase, a training set image is trained to obtain a four-layer Multi-Sensory Fusion Deep Neural Network, and a training set text is trained to obtain a three-layer text semantic network, respectively. Finally, a test image and a text are respectively mapped to an isomorphic semantic space by using two networks, so as to realize cross-media retrieval. The disclosed method can significantly improve the performance of cross-media retrieval.

Abstract: A method and apparatus for inputting data for an electronic data entry device are provided. In one embodiment, identification of an input object such as the particular fingers of a user that are used to actuate a key region is performed. The symbol associated with the actuated key region and the finger (or other input object) used is determined. In other embodiments, virtual input devices with interfaces such as QWERTY style keyboards, phone keypads, and multi-touch capable touchpads or tablets are provided in input regions. One or more video capturing devices remotely acquire actuation information from the input regions during data entry. User inputted symbols or functions are determined based on the actuations, their locations and identified input object sets that caused the actuations.

Abstract: Embodiments of the present disclosure provide a gas purifying device and an ion migration spectrometer. The gas purifying device includes a first purificant vessel, a second purificant vessel and a valve communicated between the first purificant vessel and the second purificant vessel. The valve is configured to allow a gas flows from the second purificant vessel to the first purificant vessel in a first state and to permit the gas to flow from the first purificant vessel to the second purificant vessel in a second state.

Abstract: The present invention relates to the biomedicine field, in particular to an anti-human PD-1 humanized monoclonal antibody and its applications. The invention obtains an anti-human PD-1 humanized monoclonal antibody with good specificity, high affinity and stability by screening, and the antibody can specifically bind to human PD-1 instead of binding to other members of CD28 family, block the binding of PD-L1 and PD-1 with CD80 and partially restore functions of T-cells, so it can significantly inhibit the growth of tumor.

Abstract: An intra-frame and inter-frame combined prediction method for P frames or B frames. The method comprises: self-adaptively selecting by means of a rate-distortion optimization (RDO) decision whether to use the intra-frame and inter-frame combined prediction or not; using a method for weighting an intra prediction block and an inter prediction block in the intra-frame and inter-frame combined prediction to obtain a final prediction block; and obtaining the weighting coefficient of the intra prediction block and the inter prediction block according to prediction distortion statistics of the prediction method. Therefore, prediction precision can be improved, and coding and decoding efficiency of the prediction blocks are improved.

Abstract: A coil component includes a first core having a leg portion, a second core joined to the first core with the leg portion therebetween, and a magnet disposed between the leg portion and the second core. Movement of the magnet in a first direction intersecting a direction in which the first core and the second core face each other is at least restricted by an uneven structure provided on a junction surface between the magnet and at least one of the first core and the second core.

Abstract: The first core includes a main body part extending in a first direction along a main surface of the substrate, a first foot part extending from the main body part to the second core through the substrate, and a second foot part extending from the main body part to the second core through the substrate at a position at which the coil conductor is sandwiched between itself and the first foot part in the first direction, and the insulating member includes a bottom wall part interposed between at least the first foot part and the second core, and a side wall part extending along at least either of the first foot part and the second foot part and interposed between either of the foot parts and the coil conductor.

Abstract: A self-adaptive point cloud stripe division method. The method comprises: firstly, carrying out space division with a certain depth on a point cloud to obtain a plurality of local point clouds; then, counting the number of points in each of the local point clouds, comparing same with an upper and lower limit for the number of stripe points, and determining whether the number of points satisfies a requirement; and after a series of re-segmentation or re-fusion operations on the local point clouds, adjusting the number of points in each of the local point clouds until the number of points satisfies a range, thereby obtaining a final point cloud stripe. A plurality of local structures capable of being independently coded and decoded are obtained by means of division of a point cloud stripe, and this supports parallel processing, enhances system fault tolerance, and improves coding efficiency.

Abstract: Disclosed is a method for detecting pedestrians in an image by using Gaussian penalty. Initial pedestrian boundary box is screened using a Gaussian penalty, to improve the pedestrian detection performance, especially sheltered pedestrians in an image. The method includes acquiring a training data set, a test data set and pedestrian labels of a pedestrian detection image; using the training data set for training to obtain a detection model by using a pedestrian detection method, and acquiring initial pedestrian boundary box and confidence degrees and coordinates thereof; performing Gaussian penalty on the confidence degrees of the pedestrian boundary box, to obtain confidence degree of the pedestrian boundary box after the penalty; and obtaining final pedestrian boundary boxes by screening the pedestrian boundary boxes. Thus, repeated boundary boxes of a single pedestrian are removed while reserving boundary boxes of sheltered pedestrians, thereby realizing the detection of the pedestrians in an image.

Abstract: The present invention discloses intelligent oil sludge treatment apparatuses and treatment processes.

Abstract: Embodiments of the present disclosure provide an ion mobility spectrometer device. The ion mobility spectrometer device includes: an ion mobility tube, a sampling device, and a sampling and circulating gas path. The sampling device includes a solid sample desorption device and a gas sampling device. The solid sample desorption device is configured to process the solid sample into a first mixed gas containing the solid sample, and the gas sampling device is configured to process the gas sample into a second mixed gas containing the gas sample. The sampling and circulating gas path is configured to transfer the first mixed gas and/or the second mixed gas into the ion mobility tube for detection.

Abstract: Disclosed is a deep discriminative network for person re-identification in an image or a video. Concatenation are carried out on different input images on a color channel by constructing a deep discriminative network, and an obtained splicing result is defined as an original difference space of different images. The original difference space is sent into a convolutional network. The network outputs the similarity between two input images by learning difference information in the original difference space, thereby realizing person re-identification. The features of an individual image are not learnt, and concatenation are carried out on input images on a color channel at the beginning, and difference information is learnt on an original space of the images by using a designed network. By introducing an Inception module and embedding the same into a model, the learning ability of a network can be improved, and a better differentiation effect can be achieved.

Abstract: An image feature extraction method for person re-identification includes performing person re-identification by means of aligned local descriptor extraction and graded global feature extraction; performing the aligned local descriptor extraction by processing an original image by affine transformation and performing a summation pooling operation on image block features of same regions to obtain an aligned local descriptor; reserving spatial information between inner blocks of the image for the aligned local descriptor; and performing the graded global feature extraction by grading a positioned pedestrian region block and solving a corresponding feature mean value to obtain a global feature. The method can resolve the problem of feature misalignment caused by posture changes of pedestrian, etc., and eliminate the effect of unrelated backgrounds on re-recognition, thus improving the precision and robustness of person re-identification.

Abstract: A bidirectional image-text retrieval method based on a multi-view joint embedding space includes: performing retrieval with reference to a semantic association relationship at a global level and a local level, obtaining the semantic association relationship at the global level and the local level in a frame-sentence view and a region-phrase view, and obtaining semantic association information in a global level subspace of frame and sentence in the frame-sentence view, obtaining semantic association information in a local level subspace of region and phrase in the region-phrase view, processing data by a dual-branch neural network in the two views to obtain an isomorphic feature and embedding the same in a common space, and using a constraint condition to reserve an original semantic relationship of the data during training, and merging the two semantic association relationships using multi-view merging and sorting to obtain a more accurate semantic similarity between data.

Abstract: Disclosed in the present invention is a point cloud attribution compression method based on deleting 0 elements in a quantisation matrix, including optimizing a traversal sequence for a quantisation matrix and deleting the 0 elements at the end of the data stream. The present invention may use seven types of traversal sequences at the encoding end of the point cloud attribute compression, such that the distribution of the 0 elements in the data stream may be more concentrated at the end thereof. The 0 elements at the end of the data stream may be deleted, removing redundant information and reducing the quantity of data to be entropy encoded. At the decoding end, the point cloud geometric information may be incorporated to supplement the deleted 0 elements and the quantisation matrix may be restored according to the traversal sequence, thereby improving compression performance without introducing new errors.