Abstract: A system for image quality assessment of non-aligned images includes a first deep path portion of a convolutional neural network having a set of parameters and a second deep path portion of the convolutional neural network sharing a set of parameters with the first deep path convolutional neural network. Weights are shared between the first and second deep path convolutional neural networks to support extraction of a same set of features in each neural network pathway. Non-aligned reference and distorted images are respectively provided to the first and second deep paths of the convolutional neural network for processing. A concatenation layer is connected to both the first and second deep paths convolutional neural network, and a fully connected layer is connected to the concatenation layer to receive input from both the first and second deep paths of the convolutional neural network, generating an image quality assessment as a linear regressor and outputting an image quality score.

Abstract: A system for image quality assessment of non-aligned images includes a first deep path portion of a convolutional neural network having a set of parameters and a second deep path portion of the convolutional neural network sharing a set of parameters with the first deep path convolutional neural network. Weights are shared between the first and second deep path convolutional neural networks to support extraction of a same set of features in each neural network pathway. Non-aligned reference and distorted images are respectively provided to the first and second deep paths of the convolutional neural network for processing. A concatenation layer is connected to both the first and second deep paths convolutional neural network, and a fully connected layer is connected to the concatenation layer to receive input from both the first and second deep paths of the convolutional neural network, generating an image quality assessment as a linear regressor and outputting an image quality score.

Abstract: A facial recognition method using online sparse learning includes initializing target position and scale, extracting positive and negative samples, and extracting high-dimensional Haar-like features. A sparse coding function can be used to determine sparse Haar-like features and form a sparse feature matrix, and the sparse feature matrix in turn is used to classify targets.

Abstract: A facial recognition method using online sparse learning includes initializing target position and scale, extracting positive and negative samples, and extracting high-dimensional Haar-like features. A sparse coding function can be used to determine sparse Haar-like features and form a sparse feature matrix, and the sparse feature matrix in turn is used to classify targets.

Abstract: A multi-object tracking (MOT) framework uses a dual Long Short-Term Memory (LSTM) network (Siamese) for MOT. The dual LSTM network learns metrics along with an online updating scheme for data association. The dual LSTM network fuses relevant features of trajectories to interpret both temporal and spatial components non-linearly and concurrently outputs a similarity score. An LSTM model can be initialized for each trajectory and the metric updated in an online fashion during the tracking phase. An efficient and feasible visual tracking approach using Optical Flow and affine transformations can generate robust tracklets for initialization. Thus, the MOT framework can achieve increased tracking accuracy. Further, the MOT framework has improved performance and can be flexible utilized in arbitrary scenarios.

Abstract: A facial recognition method using online sparse learning includes initializing target position and scale, extracting positive and negative samples, and extracting high-dimensional Haar-like features. A sparse coding function can be used to determine sparse Haar-like features and form a sparse feature matrix, and the sparse feature matrix in turn is used to classify targets.

Abstract: A non-hierarchical and iteratively updated tracking system includes a first module for creating an initial trajectory model for multiple targets from a set of received image detections. A second module is connected to the first module to provide identification of multiple targets using a target model, and a third module is connected to the second module to solve a joint object function and maximal condition probability for the target module. A tracklet module can update the first module trajectory module, and after convergence, output a trajectory model for multiple targets.

Abstract: A visual recognition system to process images includes a global sub-network including a convolutional layer and a first max pooling layer. A local sub-network is connected to receive data from the global sub-network, and includes at least two convolutional layers, each connected to a max pooling layer. A fusion network is connected to receive data from the local sub-network, and includes a plurality of fully connected layers that respectively determine local feature maps derived from images. A loss layer is connected to receive data from the fusion network, set filter parameters, and minimize ranking error.

Abstract: A face replacement system for replacing a target face with a source face can include a facial landmark determination model having a cascade multichannel convolutional neural network (CMC-CNN) to process both the target and the source face. A face warping module is able to warp the source face using determined facial landmarks that match the determined facial landmarks of the target face, and a face selection module is able to select a facial region of interest in the source face. An image blending module is used to blend the target face with the selected source region of interest.

Abstract: A system for image quality assessment of non-aligned images includes a first deep path portion of a convolutional neural network having a set of parameters and a second deep path portion of the convolutional neural network sharing a set of parameters with the first deep path convolutional neural network. Weights are shared between the first and second deep path convolutional neural networks to support extraction of a same set of features in each neural network pathway. Non-aligned reference and distorted images are respectively provided to the first and second deep paths of the convolutional neural network for processing. A concatenation layer is connected to both the first and second deep paths convolutional neural network, and a fully connected layer is connected to the concatenation layer to receive input from both the first and second deep paths of the convolutional neural network, generating an image quality assessment as a linear regressor and outputting an image quality score.

Abstract: A multi-object tracking (MOT) framework uses a dual Long Short-Term Memory (LSTM) network (Siamese) for MOT. The dual LSTM network learns metrics along with an online updating scheme for data association. The dual LSTM network fuses relevant features of trajectories to interpret both temporal and spatial components non-linearly and concurrently outputs a similarity score. An LSTM model can be initialized for each trajectory and the metric updated in an online fashion during the tracking phase. An efficient and feasible visual tracking approach using Optical Flow and affine transformations can generate robust tracklets for initialization. Thus, the MOT framework can achieve increased tracking accuracy. Further, the MOT framework has improved performance and can be flexible utilized in arbitrary scenarios.

Abstract: A semi-supervised model incorporates deep feature learning and pseudo label estimation into a unified framework. The deep feature learning can include multiple convolutional neural networks (CNNs). The CNNs can be trained on available training datasets, tuned using a small amount of labeled training samples, and stored as the original models. Features are then extracted for unlabeled training samples by utilizing the original models. Multi-view clustering is used to cluster features to generate pseudo labels. Then the original models are tuned by using an updated training set that includes labeled training samples and unlabeled training samples with pseudo labels. Iterations of multi-view clustering and tuning using an updated training set can continue until the updated training set is stable.

Abstract: A foreground attentive neural network is used to learn feature representations. Discriminative features are extracted from the foreground of the input images. The discriminative features are used for various visual recognition tasks such as person re-identification and multi-target tracking. A deep neural network can include a foreground attentive subnetwork, a body part subnetwork and the feature fusion subnetwork. The foreground attentive subnetwork focuses on foreground by passing each input image through an encoder and decoder network. Then, the encoded feature maps are averagely sliced and discriminately learned in the following body part subnetwork. Afterwards, the resulting feature maps are fused in the feature fusion subnetwork. The final feature vectors are then normalized on the unit sphere space and learned by following the symmetric triplet loss layer.

Abstract: The present disclosure relates to a label attaching apparatus comprising a mounting base; an attaching plate having a first end which is connected to a first position on the mounting base and a second end which is a free end and provided with a label attaching unit; and a driving cylinder including a cylinder body and a piston rod retractable and extendable relative to the cylinder body, the cylinder body being connected to a second position on the mounting base, a distal end of the piston rod opposite to the cylinder body being connected to the attaching plate between the first and second ends of the attaching plate, wherein the first end of the attaching plate is hingedly connected onto the mounting base so that an orientation of the label attaching unit is able to be changed by the movement of the piston rod.

Abstract: A vehicle having at least one camera wherein the at least one camera is configured to provide an improved view for the driver of the vehicle. The vehicle includes a body having a camera mount operably coupled therewith. The camera mount is operably coupled to the vehicle in a first technique or a second technique. The camera mount has secured to the second end thereof a camera. The camera mount is movable intermediate a first position and a second position. In the second position of the camera mount the camera secured to the second end thereof is positioned such that the camera is above the roof of the vehicle. A drive motor and transmission mechanism is provided to offer operation of the camera mount so as to move intermediate its first position and second position. A display and controller are disposed within the passenger compartment of the vehicle.

Abstract: A system for image quality assessment of non-aligned images includes a first deep path portion of a convolutional neural network having a set of parameters and a second deep path portion of the convolutional neural network sharing a set of parameters with the first deep path convolutional neural network. Weights are shared between the first and second deep path convolutional neural networks to support extraction of a same set of features in each neural network pathway. Non-aligned reference and distorted images are respectively provided to the first and second deep paths of the convolutional neural network for processing. A concatenation layer is connected to both the first and second deep paths convolutional neural network, and a fully connected layer is connected to the concatenation layer to receive input from both the first and second deep paths of the convolutional neural network, generating an image quality assessment as a linear regressor and outputting an image quality score.

Abstract: A system provides automated editing of a media file. Frames of a media file are extracted and feature vectors are generated based thereon. The feature vectors are clustered according to similarity of the content of the feature vectors and the temporal proximity of frames corresponding to the feature vectors in the media file. Clusters are ranked and selected according to the ranking. Segments of the media file corresponding to the selected clusters are combined into a highlight video. Hotspots in a media file may be identified in a media file by detecting local maxima in a measure of movement of pixels between frames. Clusters may be ranked and selected according to an iterative algorithm that identifies clusters that are the most different from an average of the feature vectors and from other selected clusters.

Abstract: Provided are a method and device for improving sound quality of stereo sound and a terminal. The method includes: an original left channel signal and an original right channel signal are acquired; phases, frequency spectrums and amplitudes of the original left channel signal and original right channel signal are acquired; a left calibrated signal is acquired according to the phase, frequency spectrum and amplitude of the original left channel signal, and a right calibrated signal is acquired according to the phase, frequency spectrum and amplitude of the original right channel signal; the left calibrated signal and the original right channel signal are superposed to generate a final right channel output signal; the right calibrated signal and the original left channel signal are superposed to generate a final left channel output signal; the final right channel output signal and the final left channel output signal are combined to form a PCM signal.

Abstract: A restrictor device is used on a cable to protect the cable from excessive bending. The restrictor device includes two or more restrictor members having a helix shape extending along a longitudinal axis X; and fasteners that connect the two or more restrictor members to each other to form the restrictor device. At least one property of the two or more restrictor members is selected such that a bending portion of the restrictor device moves along the restrictor device when a magnitude of a force applied to the restrictor device changes.

Abstract: A system for image quality assessment of non-aligned images includes a first deep path portion of a convolutional neural network having a set of parameters and a second deep path portion of the convolutional neural network sharing a set of parameters with the first deep path convolutional neural network. Weights are shared between the first and second deep path convolutional neural networks to support extraction of a same set of features in each neural network pathway. Non-aligned reference and distorted images are respectively provided to the first and second deep paths of the convolutional neural network for processing. A concatenation layer is connected to both the first and second deep paths convolutional neural network, and a fully connected layer is connected to the concatenation layer to receive input from both the first and second deep paths of the convolutional neural network, generating an image quality assessment as a linear regressor and outputting an image quality score.