22 results about "Histogram" patented technology

A method for recognizing objects in images is disclosed. The method of the present invention comprises the following steps. First, acquire a digital image. Then, select one or more objects from the image according to a certain characteristic. Next, generate an x-axis histogram and/or a y-axis histogram from the segmented image. Then, find the zeroes and maxima for the x-axis histogram and/or the y-axis histogram and use the polynomial regression analysis to determine the shape, shape and location of each of the objects in the segmented image according to the zeroes and maxima. If the two curves linking two zeroes and one maximum in the x-axis histogram and the y-axis histogram are two sloped line, the corresponding object may be determined to be a triangle. If each of the four curves linking two zeroes and two maxima is a line, the corresponding object may be determined to be a rectangle.

The invention discloses an adaptive histogram reinforced defogging method based on a dark channel. According to the method, a dark channel model is adopted to obtain a dark channel map of a foggy image, the dark channel map is divided into a plurality of square blocks, and a weighting factor of each square block is calculated; then, a sub-block histogram of each square block is subjected to statistics with an implementation method of CLAHE; according to the sub-block histogram, a sub-block limit contrast histogram is calculated; a histogram balanced grey mapping relationship table corresponding to the square block is calculated based on the contrast histogram, and contrast stretching grey mapping is calculated; a final grey mapping relationship table is calculated by using the weighting factor, the grey mapping relationship table and the contrast stretching grey mapping; by using the grey mapping tables of the four adjacent square blocks, a center pixel of each block adopts an original grey mapping relationship, while the other pixels are acquired via grey mapping interpolation of the four blocks, and a defogged image is obtained; and finally, the image is output. According to the adaptive histogram reinforced defogging method based on the dark channel, provided by the invention, the dark channel model and a traditional CLAHE image reinforcement algorithm are combined, so that depth information in a foggy day is effectively utilized and complicated transmission pattern estimation is avoided; a single image is processed; and the adaptive histogram reinforced defogging method is good in defogging effect and high in practicality.

The invention discloses a behavior recognition method based on intelligent sub-space networks. The method specifically includes the following steps of preprocessing and learning spatial characteristics, extracting spatial and temporal characteristics, representing videos as histogram vectors, training SVM classifiers and recognizing behaviors. According to the behavior recognition method based on the intelligent sub-space networks, the spatial characteristics are extracted from video data, the spatial characteristics are then pooled so that the characteristics equivalent to the spatial and temporal characteristics can be obtained, each video sequence is represented as the corresponding histogram vector about all words in a word bag through a word bag method, multiple types of behaviors are recognized through the trained SVM classifiers and a one-to-many strategy, the robustness of behavior recognition is increased, and the behavior recognition rate is increased.

The invention provides an automatic target tracking method based on color and shape features. The automatic target tracking method comprises the steps of A, inputting a video sequence and obtaining a target region of the video sequence; B, obtaining a color and shape feature joint probability distribution histogram of the target region; C, obtaining the window size and the accurate position of the target region in a current frame image according to the color and shape feature joint probability distribution histogram; D, utilizing Kalman filtering to obtain an initial searching position estimation value of the target region in a next frame image according to the window size and the accurate position of the target region in the current frame image; E, using the initial searching position estimation value as the initial size and position of a searching window of the target region in the next frame image and executing the step B so as to achieve continuous tracking of targets in the video sequence. The automatic target tracking method has the advantages of being accurate and stable in target tracking. The invention further provides an automatic target tracking system based on the color and shape features.

Computer-implemented systems and methods for visually presenting spectrum usage of optical spectrum in an optical network include displaying a network map of the optical network comprising a plurality of nodes and a plurality of links connecting the plurality of nodes to one another; responsive to obtaining spectrum data comprising channel assignments on the plurality of links and nodes as endpoints of the associated channel assignments, displaying a plurality of circular histograms to visually illustrate spectrum usage in the optical network; and adjusting the plurality of circular histograms based on selections of a plurality of endpoints in the optical network. The plurality of circular histograms visually represent the spectrum usage by representing the optical spectrum in the optical network around the associated circular histogram, and wherein each portion or segment of the plurality of circular histograms represents one of a wavelength and a portion of spectrum.

A video controller and method for performing tone-mapping of high-dynamic-range (HDR) video are provided. The video controller includes: a color-space converter, arranged to receive an input high-dynamic-range (HDR) video signal and perform a color space conversion on the input HDR video signal to obtain a first video signal having a first gamma curve; a de-gamma unit, arranged to apply a second gamma curve on the first video signal to compensate the first gamma curve to obtain a second video signal; a first histogram calculator, arranged to calculate a first histogram of a current frame of the second video signal; and a tone-mapping unit, arranged to apply a tone-mapping curve on the current frame of the second video signal according to a histogram of a previous frame of the second video signal to generate an output video signal.

The present invention discloses a method of rating the fruit according to surface colors. Red fruit samples and orange fruit samples are selected respectively to form a red fruit sample unit and an orange fruit sample unit, a fruit sample image without background is obtained, the weighted histogram of the H weight and area of the fruit sample image is calculated and then the principle component is analyzed; the principle components whose variance contribution ratio are before K are selected; when the accumulated variance contribution ratio is larger than 85 percent, the sample central points of the red fruit sample unit and the orange fruit sample unit are calculate respectively, which obtains the principle component central points of two sample units; then the Mahalanobis Distances of DMR and DMY of the red fruit sample unit and the orange fruit sample unit are calculated respectively, if DMR is less than DMY, the test object is red fruit, or else the test object is orange fruit. The present invention overcomes the problem that the practical area represented by each point on the fruit image is not completely coincident, improves the precision of fruit rating according to surface colors, and is able to rate accurately the fruit according to surface colors.

The invention provides a medical image segmentation method based on a lightweight full convolutional neural network. The method comprises the following steps: carrying out preprocessing such as graying, normalization, contrast limited adaptive histogram equalization (CLAHE) and gamma correction on a data set; randomly extracting patches from the training set and sequentially extracting patch graphs from the test set to complete data enhancement; building a full convolutional neural network architecture composed of a contraction path (left side) and an expansion path (right side), and designinga left-one-out training method for a data set with a small number of images; and finally, completing BN channel model cutting through channel sparse regularization training, cutting channels of whichscaling factors are smaller than a set threshold, finely adjusting the cut network to obtain a lightweight full convolutional neural network, and inputting test data into the network for rapid test to complete image segmentation. The lightweight full convolutional neural network not only ensures the advantage of high segmentation precision of the deep network, but also improves the test speed ofthe image segmentation network.

An imaging apparatus includes: determination means for determining whether a capture image is a frontlit image or a backlit image; detecting means for detecting a flesh-color area in the capture image; calculation means for setting weights, in a luminance histogram on the flesh-color area, on luminance of pixels included in a high-luminance area having a relatively high luminance value and luminance of pixels included in a low-luminance area having a relatively low luminance value in accordance with the determination result, and calculating modified luminance produced by modifying luminance on the flesh-color area on the basis of the weights; and correcting means for correcting brightness of the capture image on the basis of the modified luminance.

The invention discloses a fiber state detection method and system, and relates to the technical field of fiber detection. The method comprises: performing calculating on a standard fiber output light field image to obtain a standard value aperture histogram of a fiber output light field; through analyzing an acquired image of a fiber output light field to be detected, obtaining a value aperture histogram of the fiber output light field to be detected; and comparing the value aperture histogram of the fiber output light field to be detected with the standard value aperture histogram of the fiber output light field to obtain a state change of the fiber output light field to be detected. According to the invention, for the structural characteristics and optical waveguide characteristics of a multi-wrapping-layer fiber, the method is designed to realize detection of the state of the multi-wrapping-layer fiber. The method is applied to detection of fiber defects, bending states, loss parameters and the like, and can also realize real-time monitoring and feedback control of the multi-wrapping-layer fiber at the same time.

A method for improving brightness of projected images from an LED projector employing a plurality of LEDs of different colors by determining, from a histogram of a frame of an image to be projected, an effective maximum saturation. A plurality of main channels and a plurality of subchannels are created, one main channel and at least one subchannel for each color LED. Then the amplitude of the main channel and a subchannel for each color are determined based upon the effective maximum saturation of the frame of the image, followed by using the main channel and the at least one subchannel for a color to drive an LED of that color to generate the image.

An apparatus for generating a scaled image divides an image, which comprises a plurality of pixels, into a plurality of blocks by partitioning the image vertically and horizontally, whereby the color space of this image is divided into a plurality of subspaces, referred to as bins. The apparatus applies a histogram analysis, on a bin-by-bin basis, to the pixels constituting a block of interest among the plurality of blocks. Then, using the result of the histogram analysis, the apparatus decides the representative color of the block of interest in accordance with the average color of pixels belonging to a most frequent bin among the plurality of bins. A scaled image of this image is obtained by applying the above operation to all of the blocks.

The invention provides a device and a method for testing an APD measuring device. To accurately test the operating state of the APD measuring device, a test device (40) includes a test signal generator (41) that generates a test signal corresponding to the center frequency of a signal to be measured and a control unit (42) that variably controls the amplitude level of a test signal in the dynamic range of an APD measuring device (1) at random. If the test signal is input from the test device (40), the APD measuring device (1) outputs a PDF histogram manufactured according to counting of an amplitude peak value of the test signal during change of the amplitude level and a time capturing rate calculated based on the difference between the DPD data of the test signal and the referential APD data, as the test result representing a motion state.

The invention discloses a data visualization chart library system, comprising a chart drawing unit, a personalized icon style unit, a style introduction unit, a data loading and updating unit, a loading animation unit, a data dynamic updating unit, an interaction component adding unit, an event behavior unit and a mouse event processing unit which are connected with one another, wherein the chartdrawing unit is used for preparing a DOM container with height and width for the data visualization chart library system before drawing, then initializing a data visualization chart library software instance through a generate method, and transmitting the data visualization chart library software instance into an Option parameter to generate a histogram; and the personalized icon style unit is used for providing a plurality of custom configuration options and can set the style of the data graph from three levels, namely a global level, a series level and a data level. The data visualization chart library system has the advantages that the software introduction mode is simple, and the requirement of a user for functional diversification can be met.

The invention relates to a long-time correlation filtering target tracking method based on adaptive feature fusion, and belongs to the field of computer vision. The method comprises the following steps: in a target position prediction stage, respectively extracting an HOG feature and a color histogram feature of each frame of image, respectively obtaining a response diagram of each feature by using a kernel correlation filter and a color histogram model, and normalizing a peak value of the response diagram of each feature so as to dynamically adjust weights of different features; the adaptive fusion of the features is realized, and the position of the tracking target is estimated according to the fused feature response graph; two confidence coefficient detection indexes are set at the same time, and a detection filter and an SVM classifier are additionally trained and are respectively used for detecting whether a current tracking target is reliable or not, namely, whether the target is lost or not and repositioning the position of the lost target, so that long-time tracking of the target is realized.

The invention discloses a scale-adaptive target object updating method and device, and belongs to the field of computer vision. The method comprises the steps: determining an optimal image scale of atarget based on a scale pyramid; updating the target model of a next image frame of the target based on the optimal image scale when the target is not shielded. Because the scale pyramid has a largerscale range, the method can adapt to the severe changes of the target scale. The gray scale distribution features serve as the features used in the estimation of the target scale, and the features aremade into a statistical histogram according to the position and gray scale of target pixels. Because the method gives consideration to the position information of the target, the gray scale distribution features can reflect the characteristics of the target better than a grey level histogram, and can achieve the more accurate estimation of the sale through the features.