44 results about "Steerable filter" patented technology

A method and apparatus for lane detection using overhead images and positional data. A server receives positional data from a vehicle and computes a continuous trajectory. The server receives an overhead image of a road section. The server crops and processes the overhead image to remove unwanted portions. The server identifies edge features using the continuous trajectory and steerable filters. The server identifies lanes in the overhead image using a maximization algorithm, the edge filters, and the continuous trajectory.

An acoustic apparatus includes a microelectromechanical system (MEMS) device, a controlled filter coupled to the MEMS device, and an amplifier. The controllable filter and the amplifier are coupled together at a node. A cut-off frequency of the filter is selectable based upon reception or non-reception of a low frequency audio signal by the acoustic apparatus.

A listening device comprises an input transducer providing an electric input signal comprising audio and a detector coupled to the input transducer, for determining whether the electric input signal is a broadband signal or not and providing a detection signal in response. The listening device furthermore comprises a controllable filter for filtering the electric input signal being coupled to the detector and the input transducer and for outputting a filtered electric input signal such that a component of the electric input signal in the tinnitus frequency range is dampened, if the detection signal indicates that the electric input signal is a broadband signal, and outputting an unfiltered electric input signal such that a component of the electric input signal in the tinnitus frequency range is not dampened, if the detection signal indicates that the electric input signal is not a broadband signal.

An apparatus, system, method and computer program product directed to controlling corrosion of a conductive structure in contact with a corrosive environment and electrically connected to one or more anodes, wherein the anodes are less noble than the conductive structure, where the corrosion is controlled by a controllable filter and a corresponding electronic control unit configured to process at least one stored or measured parameter, and wherein the apparatus, system and method serve to prolong the lifetime of the one or more anodes by reducing, minimizing or substantially eliminating their sacrificial character.

The invention discloses an illumination unevenness image enhancement method and system. The method comprises the steps of S1, converting an illumination unevenness image to an HSV color space from an RGB color space, and extracting a brightness component image; S2, performing spatial multi-size homomorphic filtering on the brightness component image to obtain a primary filtered image; S3, performing local entropy-based steerable filtering on the primary filtered image to obtain a final filtered image; and S4, combining the final filtered image as an enhanced brightness component image with an original hue component image and an original saturation component image, and converting a combined image to the RGB color space from the HSV color space, thereby obtaining an enhanced image. According to the method and the system, the illumination unevenness image can be enhanced, so that the illumination is even, the details are rich, and the color is natural.

A communications network wherein first and second optical signals can be launched into an optical fiber in respective first and second regions of a transmission window such that controllable filters can be used to selectively recover the first and second optical signals. This partitioning of the transmission window allows two incompatible optical signals to coexist in the same optical fiber.

The invention discloses an improved dark channel prior image defogging algorithm based on fast steerable filtering, and belongs to the technical field of image defogging algorithms. The invention aimsat solving problems that the global atmosphere light intensity evaluation in a conventional dark channel single-image defogging algorithm is liable to be disturbed by a white object, the color of a sky region is distorted and the calculation is complex when a software image cutting method is used for optimizing transmissivity. According to the invention, the algorithm provided by the invention employs an improved quadtree search algorithm and the improved fast steerable filtering for refining the transmissivity, effectively improves the defogging effect, is bright in color of a recovered image, is clear and natural in visual effect, and improves the processing time to a certain degree.

A method for downsampling fluoroscopic images and enhancing guidewire visibility during coronary angioplasty includes providing a first digitized image, filtering the image with one or more steerable filters of different angular orientations, assigning a weight W and orientation O for each pixel based on the filter response for each pixel, wherein each pixel weight is assigned to a function of a maximum filter response magnitude and the pixel orientation is calculated from the angle producing the maximum filter response if the magnitude is greater than zero, wherein guidewire pixels have a higher weight than non-guidewire pixels, and downsampling the orientation and weights to calculate a second image of half the resolution of the first image, wherein the downsampling accounts for the orientation and higher weight assigned to the guidewire pixels.

The present invention relates to a method of filtering an input image for producing an output image. The method comprising the steps of, for a first image point located at first spatial coordinates:—analysing intensity levels of the input image in a first vicinity of the first image point for calculating a first direction of filtering at said first image point;—computing an intensity level of a first output image point of a first output image by applying a first steerable filter kernel to the input image, which is centered at the first image point and oriented along the calculated first direction of filtering; selecting at least a second image point in a second vicinity of said first image point in the first output image;—computing an intensity level of at least one second output image point of a second output image by applying at least a second steerable filter kernel to the input image, which is centered at the at least second image point and oriented along a second direction of filtering;—calculating a linear combination of the first output image point and of the at least second output image point for providing an output intensity level;—applying said output intensity level to an output image point of said output image, said output image point having said first spatial coordinates.

The invention relates to a remote-sensing image building location detection method based on probability. The remote-sensing image building location detection method includes the steps of (1) performing smoothing filter to acquire the image 2; (2) constructing a controllable filter; (3) convoluting; (4) binaryzating; (5) extracting structure features of twelve directions; (6) classifying straight lines and curves; (7) calculating probability of central points of candidate buildings by means of probability density functions; and (8) defining thresholds to determine location of the central points of the buildings. By determining the location of the central points of the buildings according to the probability density functions, the problem of low building-location detection efficiency caused by the conventional methods which adopt modeling and classifiers and the like is solved, accuracy and efficiency of building detection are both improved, and the effect of full automation is achieved.

An apparatus for calibrating an audio-visual (AV) signal includes a controller for generating a control signal, a controllable filter for selectively filtering the AV signal in response to the control signal to output either the AV signal or a filtered AV signal; and a calibrator for generating a group of calibrating coefficients according to the filtered AV signal and calibrating the AV signal according to the group of calibrating coefficients.

The invention relates to an image dehazing method. The method comprises the steps that A, whether dehazing processing is needed or not is judged according to the accumulated value obtained after all pixel values of dark channel graphs D of original images and the original images are subtracted respectively, if not, the original images are output and the process ends, and otherwise, step B is started; B, a global atmospheric light value is acquired according to the dark channel graphs D and the original images; C, dark channel graphs D2 of new images are calculated after each pixel of the original images is divided by the global atmospheric light value, and a white picture and the pixel values of the dark channel graphs D2 are subtracted respectively to obtain a transmittance graph T1; D, the transmittance graph T1 is refined through steerable filtering to obtain a transmission rate graph T2; E, dehazed images are output through a haze graph formation model. The image dehazing method can effectively improve the image dehazing effect, simplify the dehazing processing process, greatly shorten image processing time and effectively improve the image processing efficiency.

The application relates to a controllable filter (1) that is placed between an antenna on the one side and a receiver and transmitter on the other, where the receiver branch can be switched between two filter characteristics. In this way, a simple device for allowing simplex reception in the transmission band can be provided. This filter is preferably applied as a frontend filter (FP) in the fixed part of a cordless telephone system and/or in a mobile station to thereby provide said mobile station with walkie-talkie capabilities.

A self-tracking bandpass filter (1) comprising a controllable filter (10), the controllable filter (10) having a signal input (11) for receiving a signal (φ) to be filtered, for providing a bandpass output of a bandpass signal (SBP) at a center frequency (f0), and a control input (15) for receiving a control signal (SC); and a control input (15) coupled to a controllable filter (10) A control unit (20) for generating such a control signal (SC) that the center frequency (fO) of the bandpass characteristic (HB) of the controllable filter (10) is substantially equal to that of the input signal (φ) frequency (f).

The present invention is CW pulse narrow band tracking filter and its tracking filter method. The tracking filter includes one controllable filter circuit, one A/D converter and one signal processing circuit. In the signal processing circuit, signal is acquired, filtered digitally and resolved to obtain the frequency and period time of the signal; and resolved frequency is then used in controlling the central frequency and bandwidth of the controllable filter circuit, so as to realize the tracking filtering on CW pulse. The present invention is superior to traditional tracking filter, which tends to track noise automatically in the pulse signal interval to loss lock on signal. The present invention adopts variable bandwidth frequency hopping search mode and sampling based on echo time, and thus has reduced operation amount and capacity of performing tracking filter in low S/N ratio. The present invention is suitable for tracking filter on both CW pulse signal and continuous signal.

The invention discloses an illumination-insensitive fast stereo matching method based on a local linear model. It is hard for a mainstream method in the prior art to realize accurate matching under the condition of illumination inconsistency; before matching, color correction of a stereo image is carried out first; and however, it is very difficult to obtain a color-consistent stereo image, and still no algorithm can finish accurate stereo image correction at present. The method is characterized by converting an RGB color space of an image into a CIE-Lab color space; based on the local linear model, eliminating influence of illumination inconsistency on stereo image left-right view matching; carrying out processing through an improved NCC matching algorithm based on steerable filtering; and finally, carrying out stereo disparity image solving on matching cost. The method can carry out matching no longer on dependence of the color information in stereo image left and right images in the stereo matching process, so that operating efficiency and matching accuracy are high, and the problem that it is hard for stereo matching to realize accurate matching under the condition of illumination inconsistency is effectively solved.

An apparatus, system, method and computer program product directed to controlling corrosion, particularly space weather induced corrosion, of a conductive structure in contact with a corrosive environment and coated with a semiconductive coating, where the corrosion is controlled by a controllable filter and a corresponding electronic control unit configured to process and adjust the controllable filter in response to at least one measured parameter associated with space weather effects on the conductive structure.

A computer-implemented method for training a scale-equivariant convolutional neural network. The scale-equivariant convolutional neural network is configured to determine an output signal characterizing a classification of an input image of the scale-equivariant convolutional neural network. The scale-equivariant convolutional neural network includes a convolutional layer. The convolutional layer is configured to provide a convolution output based on a plurality of steerable filters of the convolutional layer and a convolution input. The convolution input is based on the input image and the steerable filters are determined based on a plurality of basis filters. The method for training includes training the plurality of basis filters.

The present invention relates to a node (1) in a wireless communication network, where the node (1) is arranged to receive at least two signals (y₁(n), y₂(n)) which correspond to at least two transmitted uncorrelated signal streams (x₁(n), X₂(n)) which have been transmitted to the node via a channel. The channel (2) is signals represented by means of a channel matrix (H(q⁻¹)). The node (1) comprises a controllable filter structure (3) that is arranged to diagonalize the channel matrix (H(q⁻¹)) such that a channel matrix estimation (Ĥ(q⁻¹)) may be obtained. The present invention also relates to a corresponding method.