1837 results about "Feature detection" patented technology

The present invention is a system and method for detecting facial features of humans in a continuous video and superimposing virtual objects onto the features automatically and dynamically in real-time. The suggested system is named Facial Enhancement Technology (FET). The FET system consists of three major modules, initialization module, facial feature detection module, and superimposition module. Each module requires demanding processing time and resources by nature, but the FET system integrates these modules in such a way that real time processing is possible. The users can interact with the system and select the objects on the screen. The superimposed image moves along with the user's random motion dynamically. The FET system enables the user to experience something that was not possible before by augmenting the person's facial images. The hardware of the FET system comprises the continuous image-capturing device, image processing and controlling system, and output display system.

Detecting video phenomena, such as fire in an aircraft cargo bay, includes receiving a plurality of video images from a plurality of sources, compensating the images to provide enhanced images, extracting features from the enhanced images, and combining the features from the plurality of sources to detect the video phenomena. Extracting features may include determining an energy indicator for each of a subset of the plurality of frames. Detecting video phenomena may also include comparing energy indicators for each of the subset of the plurality of frames to a reference frame. The reference frame corresponds to a video frame taken when no fire is present, video frame immediately preceding each of the subset of the plurality of frames, or a video frame immediately preceding a frame that is immediately preceding each of the subset of the plurality of frames. Image-based and non-image based techniques are described herein in connection with fire detection and/or verification and other applications.

A pattern detecting apparatus has a plurality of hierarchized neuron elements to detect a predetermined pattern included in input patterns. Pulse signals output from the plurality of neuron elements are given specific delays by synapse circuits associated with the individual elements. This makes it possible to transmit the pulse signals to the neuron elements of the succeeding layer through a common bus line so that they can be identified on a time base. The neuron elements of the succeeding layer output the pulse signals at output levels based on a arrival time pattern of the plurality of pulse signals received from the plurality of neuron elements of the preceding layer within a predetermined time window. Thus, the reliability of pattern detection can be improved, and the number of wires interconnecting the elements can be reduced by the use of the common bus line, leading to a small scale of circuit and reduced power consumption.

Methods and systems for automatically generating a mask delineating a region of interest (ROI) within an image containing skin are disclosed. The image may be of an anatomical area containing skin, such as the face, neck, chest, shoulders, arms or hands, among others, or may be of portions of such areas, such as the cheek, forehead, or nose, among others. The mask that is generated is based on the locations of anatomical features or landmarks in the image, such as the eyes, nose, eyebrows and lips, which can vary from subject to subject and image to image. As such, masks can be adapted to individual subjects and to different images of the same subjects, while delineating anatomically standardized ROIs, thereby facilitating standardized, reproducible skin analysis over multiple subjects and/or over multiple images of each subject. Moreover, the masks can be limited to skin regions that include uniformly illuminated portions of skin while excluding skin regions in shadow or hot-spot areas that would otherwise provide erroneous feature analysis results. Methods and systems are also disclosed for automatically registering a skin mask delineating a skin ROI in a first image captured in one imaging modality (e.g., standard white light, UV light, polarized light, multi-spectral absorption or fluorescence imaging, etc.) onto a second image of the ROI captured in the same or another imaging modality. Such registration can be done using linear as well as non-linear spatial transformation techniques.

The invention provides a text emotion classification method based on the joint deep learning model which relates to the text emotion classification method. The method is designed with the object of solving the problems with the dimension disaster and sparse data incurred from the existing support vector machine and other shallow layer classification methods. The method comprises: 1) processing each word in the text data; using the word2vec tool to train each processed word in the text data so as to obtain a word vector dictionary; 2) obtaining the matrix M of each sentence; training the matrix M by the LSTM layer and converting it into vector with fixed dimensions; improving the input layer; generating d-dimensional h word vectors with context semantic relations; 3) using a CNN as a trainable characteristic detector to extract characteristics from the d-dimensional h word vectors with context semantic relations; and 4) connecting the extracted characteristics in order; outputting to obtain the probability of each classification wherein the classification with the maximal probability value is the predicated classification. The invention is applied to the natural language processing field.

Locating an eye includes generating an intensity response map by applying a 3-rectangle filter and applying K-mean clustering to the map to determine the eye. Locating an eye corner includes applying logarithm transform and grayscale stretching to generate a grayscale eye patch, generating a binary map of the patch by using a threshold based on a histogram of the patch, and estimating the eye corner by averaging coordinates weighted by minimal eigenvalues of spatial gradient matrices in a search region based on the binary map. Locating a mouth corner includes generating another intensity response map and generating another binary map using another threshold based on another histogram of the intensity response map. Locating a chin or a cheek includes applying angle constrained gradient analysis to reject locations that cannot be the chin or cheek. Locating a cheek further includes removing falsely detected cheeks by parabola fitting curves through the cheeks.

A system and method for automatically determining if a remote client is a human or a computer. A set of HIP design guidelines which are important to ensure the security and usability of a HIP system are described. Furthermore, one embodiment of this new HIP system and method is based on human face and facial feature detection. Because human face is the most familiar object to all human users the embodiment of the invention employing a face is possibly the most universal HIP system so far.

Apparatus and methods for detecting salient features. In one implementation, an image processing apparatus utilizes latency coding and a spiking neuron network to encode image brightness into spike latency. The spike latency is compared to a saliency window in order to detect early responding neurons. Salient features of the image are associated with the early responding neurons. A dedicated inhibitory neuron receives salient feature indication and provides inhibitory signal to the remaining neurons within the network. The inhibition signal reduces probability of responses by the remaining neurons thereby facilitating salient feature detection within the image by the network. Salient feature detection can be used for example for image compression, background removal and content distribution.

The invention discloses a real-time panoramic image stitching method of aerial videos shot by an unmanned plane. The method comprises the steps of: utilizing a video acquisition card to acquire images which are transmitted to a base station in real time by an unmanned plane through microwave channels, carrying out key frame selection on an image sequence, and carrying out image enhancement on key frames; in the image splicing process, firstly carrying out characteristic detection and interframe matching on image frames by adopting an SURF (speeded up robust features) detection method with good robustness; then reducing the series-multiplication accumulative errors of images in a frame-to-mosaic image transformation mode, determining images which are not adjacent in time sequence but adjacent in space on a flight path according to the GPS (global positioning system) position information of the unmanned plane, optimizing the frame-to-mosaic transformation relation, determining image overlapping areas, thereby realizing image fusion and the panoramic image construction and realizing real-time effect of carrying out flying and stitching simultaneously; and in image transformation, based on adjacent frame information in a vision field and adjacent frame information in airspace, optimizing image transformation to obtain the accurate panoramic images. The stitching method has good real-time performance, is fast and accurate and meets the requirements of application occasions in multiple fields.

A TCP/IP-based software application client/server communication channel construction mechanism is presented. The mechanism constructs a dedicated bi-directional data transmission for a single client/server session. The server software, running on a server host computer, provides session-based data services to one or more software clients. The client process supports one or more execution environments including: text-based scripting, Java or ActiveX components. The invoked code executing within the client computer detects the most attractive execution environment available within its execution environments, and uses the services of that execution environment client process to establish an optimal communication channel back to the server software process. Depending on the execution environment(s) available to the invoked code, the communication channel may be: (a) a content-independent, bi-directional transmission control protocol (TCP) connection, or (b) a content-dependent TCP connection, such as a TCP connection supporting only a single HTTP request/response pair.

Disclosed in the invention is an improved method of a RGB-D-based simultaneously localization and mapping (SLAM) algorithm. The method comprises two parts: a front-end part and a rear-end part. The front-end part is as follows: feature detection and descriptor extraction, feature matching, motion conversion estimation, and motion conversion optimization. And the rear-end part is as follows: a 6-D motion conversion relation initialization pose graph obtained by the front-end part is used for carrying out closed-loop detection to add a closed-loop constraint condition; a non-linear error function optimization method is used for carrying out pose graph optimization to obtain a global optimal camera pose and a camera motion track; and three-dimensional environment reconstruction is carried out. According to the invention, the feature detection and descriptor extraction are carried out by using an ORB method and feature points with illegal depth information are filtered; bidirectional feature matching is carried out by using a FLANN-based KNN method and a matching result is optimized by using homography matrix conversion; a precise inliners matching point pair is obtained by using an improved RANSAC motion conversion estimation method; and the speed and precision of point cloud registration are improved by using a GICP-based motion conversion optimization method.

A video surveillance system includes multiple video cameras. The surveillance system is configured with an arrangement to separate the surveillance functions and assign different surveillance functions to different cameras. A master camera is assigned the surveillance of large area surveillance and tracking of object movement while one or more slave cameras are provided to dynamically rotate and adjust focus to obtain clear image of the moving objects as detected by the master camera. Algorithms to adjust the focus-of-attention are disclosed to effectively carry out the tasks by a slave camera under the command of a master camera to obtain images of a moving object with clear feature detections.

A method of detecting anomalies in a communication system includes: providing a first packet flow portion and a second packet flow portion; extracting samples of a symbolic packet feature associated with a traffic status of the first and second packet flow portions; computing from the extracted samples a first statistical concentration quantity and a second statistical concentration quantity of the symbolic feature associated with the first and second packet flow portions, respectively; computing from the concentration quantities a variation quantity representing a concentration change from the first packet flow portion to the second packet flow portion; comparing the variation quantity with a comparison value; and detecting an anomaly in the system in response to the comparison.

An apparatus, system, and method are disclosed for detecting seizure symptoms in an individual 102. A sensor module 202 receives physiological data for an individual 102 from one or more sensors 110, 112, such as a heart activity sensor. A feature detection module 204 detects a predefined feature 500, 520, 530, 540 in the physiological data. The predefined feature 500, 520, 530, 540 is associated with a seizure or another medical condition. An alert module 206 broadcasts an alert in response to the feature detection module 204 detecting the predefined feature 500, 520, 530, 540.

Provide are systems methods and devices for diagnosing disease in medical images. In certain aspects, disclosed is a method for training a neural network to detect features in a retinal image including the steps of: a) extracting one or more features images from a Train_0 set, a Test_0 set, a Train_1 set and a Test_1 set; b) combining and randomizing the feature images from Train_0 and Train_1 into a Training data set; c) combining and randomizing the feature images from Test_0 and Test_1 into a testing dataset; d) training a plurality of neural networks having different architectures using a subset of the training dataset while testing on a subset of the testing dataset; e) identifying the best neural network based on each of the plurality of neural networks performance on the testing data set; f) inputting images from Test_0, Train_1, Train_0 and Test_1 to the best neural network and identifying a limited number of false positives and false negative and adding the false positives and false negatives to the training dataset and testing dataset; and g) repeating steps d)-g) until an objective performance threshold is reached.

The invention discloses a method for automatic correction and tiled display of plug-and-play large screen projections. The method comprises the steps as follows: adaptively generating a checker pattern with a certain resolution ratio, projecting by projectors in sequence, capturing by cameras, carrying out feature detection and identification on complicated projection surfaces or checkers under the illumination condition with a multi-feature detection method based on color and geometry, and creating a Bezier curve function to represent a corresponding relation of points between projector images and camera images; and obtaining effective display areas for screen projection with a quick approach method, determining a corresponding relation of the projection contents of the projectors and the display areas, carrying out geometrical distortion on the images to be projected for geometrical correction, and calculating weight values of pixels in projection overlapping areas with a distance-based nonlinear weight value distribution method for fusion of edge brightness. According to the invention, a plurality of projection images of irregular surfaces can be aligned and seamlessly spliced, and the whole method is simple and easy to use, higher in autonomy and better in seamless splicing performance.

A method of matching image features with reference features comprises the steps of providing a current image, providing a set of reference features, wherein each of the reference features comprises at least one first parameter which is at least partially indicative of a position and/or orientation of the reference feature with respect to a global coordinate system, wherein the global coordinate system is an earth coordinate system or an object coordinate system, or at least partially indicative of a position of the reference feature with respect to an altitude, detecting at least one feature in the current image in a feature detection process, associating with the detected feature at least one second parameter which is at least partially indicative of a position and/or orientation of the detected feature, or which is at least partially indicative of a position of the detected feature with respect to an altitude, and matching the detected feature with a reference feature by determining a similarity measure.

Sensory input processing apparatus and methods useful for adaptive encoding and decoding of features. In one embodiment, the apparatus receives an input frame having a representation of the object feature, generates a sequence of sub-frames that are displaced from one another (and correspond to different areas within the frame), and encodes the sub-frame sequence into groups of pulses. The patterns of pulses are directed via transmission channels to detection apparatus configured to generate an output pulse upon detecting a predetermined pattern within received groups of pulses that is associated with the feature. Upon detecting a particular pattern, the detection apparatus provides feedback to the displacement module in order to optimize sub-frame displacement for detecting the feature of interest. In another embodiment, the detections apparatus elevates its sensitivity (and/or channel characteristics) to that particular pulse pattern when processing subsequent pulse group inputs, thereby increasing the likelihood of feature detection.

An image processing apparatus includes an image acquisition unit configured to acquire a tomographic image of an object, a layer detection unit configured to detect a layer that constitutes the object from the tomographic image acquired by the image acquisition unit, and a normal structure estimation unit configured to estimate a normal structure of the layer based on the layer detected by the layer detection unit and a feature that is modified by a lesion of the layer.

Detecting video phenomena, such as fire in an aircraft cargo bay, includes receiving a plurality of video images from a plurality of sources, compensating the images to provide enhanced images, extracting features from the enhanced images, and combining the features from the plurality of sources to detect the video phenomena. The plurality of sources may include cameras having a sensitivity of between 400 nm and 1000 nm and/or may include cameras having a sensitivity of between 7 and 14 micrometers. Extracting features may include determining an energy indicator for each of a subset of the plurality of frames. Detecting video phenomena may also include comparing energy indicators for each of the subset of the plurality of frames to a reference frame. The reference frame corresponds to a video frame taken when no fire is present, video frame immediately preceding each of the subset of the plurality of frames, or a video frame immediately preceding a frame that is immediately preceding each of the subset of the plurality of frames.