89 results about "Image tracing" patented technology

The invention discloses a deep counterfeit face image source tracing evidence obtaining method and system based on a diffusion model, and relates to the technical field of deep counterfeit face image evidence obtaining. The method comprises the following steps: constructing a robust watermark framework based on a diffusion model, wherein the framework comprises a diffusion model encoder, a watermark decoder and a freeze parameter auto-encoder; in a training stage, a face image and a watermark are constructed as conditions, a diffusion model encoder is guided to denoise, and a corresponding watermark image is predicted; the trained robust watermark framework based on the diffusion model is used for testing, a specific depth forging model is added in the sampling stage for guiding, and a more robust watermark image is generated through iterative denoising; in the evidence obtaining stage, the watermark in the forged image to be subjected to evidence obtaining is obtained through the watermark decoder. The method can effectively cope with different counterfeiting technologies, provides a more robust and reliable deep counterfeiting traceability evidence obtaining scheme, and has strong adaptability and wide application prospects.

Disclosed are approaches that may provide image-guidance to interventionalists by providing true 3D visualization and quantitative feedback in real-time. A guidance system may allow a physician to manipulate a medical device and see a 3D rendering with quantitative feedback floating in mixed reality, next to standard monitors. Image tracking may detect and co-register the medical device's 3D position using, for example, bi-plane C-arm X-ray fluoroscopy and provide a 3D trajectory as quantitative feedback. Patterns in a fluoroscopic image may be used to accurately determine an object's z-position from a single angle projection.

Aspects of this technical solution can include identifying, by a processor coupled to non-transitory memory, a plurality of bounding boxes for one or more objects depicted in each image of a sequence of images captured during operation of an autonomous vehicle, allocating, by the processor and based on corresponding positions of the bounding boxes in each image and corresponding time stamps, one or more of the bounding boxes to one or more tracking identifiers each indicating trajectories of corresponding objects, generating, by the processor and based on the time stamps and the bounding boxes allocated to each of the tracking identifiers, one or more tracking images for each of the tracking identifiers, each of the tracking images including visual indications of the time stamps, and training, by the processor and based on the tracking images, an artificial intelligence model to output an indication of a type of trajectory.

The invention discloses an image watermarking method for a stable diffusion model, and the method comprises the steps: selecting one of all potential features in a reverse diffusion process of the stable diffusion model as a target potential feature, and constructing an image watermark generation recognition model which comprises a watermark information embedding module and a watermark extraction module; the watermark information embedding module is used for embedding user watermark information into the target potential features, and inputting the obtained watermark target potential features instead of the original target potential features into the remaining sub-models of the stable diffusion model to generate a watermark image; and the watermark information extraction module is used for extracting user watermark information from the watermark potential features obtained by carrying out DDIM inversion on the watermark image, and the user generating the watermark image can be identified according to the extracted user watermark information. According to the method, the user watermark information is embedded into the potential features of the stable diffusion model, so that the image traceability with high robustness and high concealment is realized on the premise of not damaging the quality of the generated image.

The invention relates to the technical field of image tracking algorithms, in particular to a target image tracking algorithm based on a Transform architecture. A template and an image of a search area are segmented into standard sizes and connected into a cascade vector, ViT is adopted as a backbone of a proposed framework, image pairs of a target and the search area are used as input, feature extraction and feature fusion are simultaneously carried out through a structured prior Transform network module, and feature extraction and feature fusion are carried out at the same time. A self-attention spatial prior generation network (SSP) is used for simulating spatial feature association, and a cross-attention-based Gaussian prior (CGP) and a random prior module (CRP) are proposed and are respectively used for establishing a spatial dependency relationship between feature marks and a matching relationship between a target and a search region. And finally, taking the output of the structured prior Transform network module as the input of the BBox prediction network to generate a tracking result. Therefore, better visual target tracking performance is realized.

The invention belongs to the technical field of intelligent traffic, particularly relates to a tunnel global tracking method based on an old camera re-identification technology, and aims to accurately realize global tracking and trajectory generation of a tunnel traffic target in a low-cost manner. The method comprises the steps of receiving an RSTP data stream, processing the RSTP data stream through a detection algorithm to obtain a sensing result, and feeding back the sensing result to the image tracking module. And processing a time sequence sensing result, adding vehicle speed information and a vehicle ID, and generating a tracking result including the vehicle ID, a target frame, a vehicle speed, a vehicle type, a color, a timestamp, a base station number and a camera number. On the basis of vehicle pixel picture data, upstream sample library candidates are matched for a current point position sample library target, if matching succeeds, the target inherits the global ID, if matching fails, the point position ID is reserved, and the target is output to a downstream point position and a simulation platform after the sample library is updated. And the trajectory generation post-processing module constructs a target matching relationship between adjacent camera sections through a frame insertion mode according to a feature re-identification result, and generates a tunnel global complete trajectory chain.

To reduce the possibility that a person who first entered a passageway after authentication may pass through the passageway when a person who was authenticated before entering the passageway and the person who first entered the passageway after authentication are different individuals, at a ticket gate, etc. at train stations.SOLUTION: In a passage control system according to the embodiment, a control device performs the processing of: inputting authentication information for a first person before entering a passageway; verifying whether the first person is an authorized passerby based on the authentication information; detecting the first person's face based on a captured image when the first person is verified as the authorized passerby; tracking the first person's face based on multiple captured images in sequence; setting an image region based on a detected position of a second person who first entered the passageway after the first person was verified as the authorized user; determining whether the face of the tracked first person is within the image region in the captured image when the image region is set; and performing control to close an opening / closing door when determined to be not within the image region.SELECTED DRAWING: Figure 7

An image tracking processing method, comprising: obtaining current frame point cloud data; preprocessing the current frame point cloud data to generate a projected image; obtaining a standard region image corresponding to standard frame point cloud data; calling a target tracking model, and obtaining a candidate region label corresponding to a candidate region based on the projected image and the standard region image; and determining a target tracking region corresponding to the current frame point cloud data according to the candidate region label.

The images shown in the diagram are interactive graphical user interfaces (GUIs) that visually communicate to the user information related to (i) unit operations in manufacturing processes such as cell and gene therapy manufacturing processes and the creation and / or modification and / or analysis of data generated therefrom, (ii) combinations of live data and textual information from databases, and (iii) interactive data analysis (e.g., patient data visualization systems) and systems for analyzing data related to manufacturing processes. For example, in one embodiment, the images shown in the diagram provide the user with multiple levels of detail about the experiment and manufacturing process, and such details can be visually presented to the user in various forms including specific combinations of charts and / or dynamic elements, which can be manipulated by the user in an interactive manner. (C. Data Display and Interactive Images) (Cell Data Images) In one embodiment, the data display image includes cell data images that track changes in various metrics (e.g., performance metrics) for multiple cell lines over the course of the manufacturing and / or experimental process. The cell data images can be presented as an organized dashboard, as shown in the diagram. In one embodiment, an interactive dashboard for viewing cell data displays data for one or more lots. In another embodiment, data for multiple lots may be displayed and presented visually to the user. In yet another embodiment, a specific lot may be automatically viewed.

The present invention discloses a method and apparatus for image provenance tracing, comprising: training a provenance graph construction model based on information about modification relationship identification and modification direction judgment of feature graphs of input image pairs; initializing a known basic database to obtain a relational database containing modification relationships between data stored in a graph structure and basic database data; performing global feature extraction on target images for image provenance tracing to obtain a set of candidate images; and constructing a directed provenance graph for the candidate image set based on the provenance graph construction model, performing image modification relationship backtracing on the candidate image set. The present invention can accurately analyze modification relationships between images to construct a provenance graph, providing a precise and efficient image provenance implementation solution suitable for large-scale scenarios.

Systems and methods for panoramic imaging in 2D and 3D ultrasound imaging are disclosed. The techniques disclosed herein use image registration to combine the 2D images into a 3D volume. Ultrasound images collected in different locations are combined using image tracking. When the transducer moves along a lateral direction (in line with a longitudinal axis of the transducer), a portion of the image is overlapped between frames and the overlapping frames can be patched. When the transducer moves along an elevational direction (non-parallel to the longitudinal axis of the transducer), there is no overlap between frames and the non-overlapping frames are not patched together but are stored side-by-side to create a volume. The storage and patching processes can be adaptive (e.g., the images are not stored or patched if there is minimal difference between neighboring frames).

The invention relates to a mirror image tracing method and device, a computer program product and electronic equipment, and relates to the technical field of computers, and the method comprises the steps: obtaining a container mirror image and a historical construction command corresponding to the container mirror image, generating mirror image construction historical record data based on the historical construction command, and storing the mirror image construction historical record data in a database; generating mirror image historical data according to the mirror image construction historical record data; matching mirror image historical data of a first container mirror image with mirror image historical data of the container mirror images to obtain a parent mirror image of the first container mirror image in the container mirror images; and storing the first container mirror image and a parent mirror image of the first container mirror image in a mirror image traceability table, and performing bidirectional traceability based on the mirror image traceability table to obtain a traceability result of the container mirror image to be traced. According to the invention, bidirectional traceability of the container mirror image is realized, and traceability precision is improved.

The present disclosure relates to systems, non-transitory computer-readable media, and methods for detecting and tracking edges in raster images using an advanced edge detection algorithm. For example, the disclosed systems generate a local histogram ranking of pixels within a sliding pixel window in a raster image, based on pixel values ​​located within the sliding pixel window. In some embodiments, the disclosed systems determine an edge thickness for an edge depicted in a region of the raster image enclosed by the sliding pixel window by comparing pixel ranks specified by the local histogram ranking of the pixels.In certain embodiments, the disclosed systems also provide the edge for display based on determining the edge thickness for the area of ​​the raster image enclosed by the sliding pixel window.

The present disclosure provides a method and apparatus for determining a pose of a tracked object in an image tracking process. In this method, a current frame of image and first point cloud information corresponding to the current frame of image are acquired, the first point cloud information being based on a relative relationship between an image capture device and an ambient space: and a pose of the tracked object in the current frame of image relative to the image capture device is determined according to the first point cloud information and object feature information of the tracked object if the tracked object is in a static state, the object feature information being based on a relative relationship between the tracked object and the ambient space. In this method, by using the first point cloud information to convert the relative relationship of the tracked object, it facilitates determining the pose of the tracked object.

This application provides a method, apparatus, device, and storage medium for target selection in adaptive cruise control, applicable to the field of adaptive cruise control technology. The method includes: acquiring image detection data of the area in front of the vehicle using an onboard camera, and acquiring an image tracking target and the vehicle's current lane based on the image detection data; acquiring radar detection data of the area in front of the vehicle using an onboard radar, and acquiring a radar tracking target based on the radar detection data; fusing the image tracking target and the radar tracking target to obtain a fused target; selecting the fused target that continuously meets target selection criteria within a preset time period as a potential following target; wherein the target selection criteria include pixel-level selection criteria and world-level selection criteria; and selecting the potential following target with the smallest longitudinal distance to the vehicle as the cruise following target. This application's method, by combining pixel-level selection and world-level selection, effectively improves the accuracy of adaptive cruise control target selection.

The invention relates to the technical field of space debris orbit forecasting, in particular to a space debris orbit forecasting correction method based on machine learning and triangular characteristic decomposition. The azimuth angle and the pitch angle are converted into six-dimensional feature vectors through triangular feature decomposition, and a real-time mapping model of orbit forecast deviation is established in combination with a neural network. The two bottleneck problems on which traditional image tracking depends are solved: 1) in a strong laser interference scene, tracking failure caused by background light saturation of an optical camera is avoided, and real-time image feedback is replaced by deviation prediction driven by historical echo data; 2) when space debris in an earth shadow area is invisible, constructing a training set by using deviation data captured at the moment of successful laser ranging, and realizing autonomous correction of forecast errors under a non-illumination condition; the method gets rid of dependence on space debris optical visibility, realizes all-day and all-arc telescope pointing deviation correction, and improves the reliability and continuity of space debris orbit prediction under complex working conditions.

This application discloses an image tracking processing method and apparatus. The method includes: obtaining feature points and attributes of each frame image in a real-time video; performing a geometric transformation on a reference frame image according to preset transformation conditions to obtain a relation table; determining a lookup index for at least one point pair based on the attribute differences between the first and second feature points of multiple point pairs; determining an evaluation value for each lookup index based on the template image corresponding to each lookup index in the relation table; obtaining the target rotation angle and target scaling scale based on the evaluation value of each lookup index; and determining the target transformation matrix based on the target rotation angle and target scaling scale. This method can establish a relation table by performing a geometric transformation on the reference frame image, quickly calculate the evaluation values ​​of multiple lookup indices through the relation table, and use multiple evaluation values ​​to quickly find the target rotation angle and target scaling scale, effectively improving the processing efficiency of image tracking.

Provided is an information processing device comprising a processor and a communication device. The communication device receives a captured image of a person passing through a prescribed passageway. The processor uses the captured image to track the person, and when the person is positioned in a verification zone that is set on the entrance side of the passageway, the processor uses the captured image of the person to verify the person. When a person for whom verification has failed is positioned in a passage determination zone set on the exit side of the passageway, the processor displays, on a prescribed monitor, information indicating that the verification has failed. When a person for whom verification has failed has exited to the exit side from the passage determination zone, the processor displays, on the monitor, information indicating that said person has made unauthorized passage.

A system includes a sensor, a weight sensor, and a tracking subsystem. The tracking subsystem receives an image feed of top-view images generated by the sensor and weight measurements from the weight sensor. The tracking subsystem detects an event associated with an item being removed from a rack in which the weight sensor is installed. The tracking subsystem determines that a first person or a second person may be associated with the event. In response to determining that the first or second person may be associated with the event, buffer frames are stored of top-view images generated by the sensor during a time period associated with the event. The tracking subsystem then determines, using at least one of the stored buffer frames and a first action-detection algorithm, whether an action associated with the event was performed by the first person or the second person.

The present application discloses a deep discriminant target tracking method, device, electronic device and computer-readable storage medium for improving the robustness and effectiveness of target tracking. The method includes: determining the search area of ​​the current frame image based on the center position of the target area of ​​the previous frame image tracking; dividing the search area into components to obtain a first overall component and a first local component; based on the features of the first overall component and the features of each first local component, using the corresponding correlation filter of each component to act on each component respectively, to obtain the correlation response map of the first overall component and the correlation response map of each first local component; based on the correlation response map of the first overall component and the correlation response map of each first local component, to obtain the overall correlation response map; using a particle filter framework to collect multiple candidate tracking target areas from the overall correlation response map, and using the candidate tracking target area with the highest confidence as the tracking target area of ​​the current frame image.

A method for realizing tapping or touch functionality with realistic touch sensation, applied in a system with XR augmented reality wearables and head-mounted displays. Preset points are marked on the joint lines of the palm. The user can see the functional areas bound to these preset points through the glasses. Two trigger points WL and WR are set. The system acquires N image video streams with parallax information. For the N images of the same time sequence, the position of the trigger fingertip P in each image is tracked and evaluated to determine whether it lies between the two trigger points WL and WR of any functional area. The x-coordinate values ​​of the three target points WL, P, and WR are acquired. The differences between WL and P, and between P and WR, are calculated, and their ratios are determined.Only when all ratios in the N images match is it determined that the trigger fingertip P has touched the functional area, and the corresponding content of the functional area is output or activated. This invention makes it possible to precisely determine whether a real touch has occurred through purely visual calculations. This involves touching the palm of the hand or an object surface, providing a realistic touch sensation when typing or using touch functions.

Provided is a tracking device comprising: a division processing unit that, for each frame image of a frame image group including a plurality of frame images, on the basis of a tracking result obtained in any other frame image followed by or following the frame image, combines a result of tracking an object in an inference image selected from among a plurality of divided images obtained by dividing the frame image with a result of tracking an object in a downsampled image other than the inference image among the plurality of divided images; an overall processing unit that, for each of the frame images, tracks an object in a reduced image obtained by reducing the frame image; and a synthesis unit that, for each of the frame images, outputs, as a tracking result of the frame image, overlapping tracking results among tracking results obtained by synthesizing a tracking result from the division processing unit and a tracking result from the overall processing unit with an excessive tracking result removed.

The present application discloses an irregular image stroking method, device, equipment and storage medium, which relates to the field of image processing technology, including: judging whether the image to be stroked has an alpha channel, determining a grayscale image based on the judgment result, and binarizing the grayscale image using an adaptive threshold algorithm; searching for image contours in the binary image based on a preset contour search algorithm, and screening the obtained closed contours; sampling the screened contours and optimizing each sampling point, and obtaining a linked list based on the optimized sampling points less than a preset distance threshold; performing vector unitization and dot product operations on the target vector constructed based on the optimized sampling points in the linked list, and deleting the optimized sampling points through the dot product value; adding the remaining sampling points of the local contour in the screened contour, and fitting the target sampling points using a preset parameterized curve algorithm to obtain a target stroked image. The smoothness and accuracy of image stroking can be improved.

The present invention discloses a multi-scale method and device for target tracking in UAV aerial photography, which relates to the technical field of image tracking and includes: obtaining an aerial photography video of a UAV, inputting the initial frame and the current frame of the aerial photography video of the UAV into the template branch and the search branch of a Siamese tracking network constructed based on the G-ResNet network, respectively outputting three groups of first weighted feature maps and second weighted feature maps from three convolutional blocks of layer2, layer3, and layer4 of the G-ResNet network, and using a plurality of anchor-free region proposal networks to perform weighted fusion on the three groups of first weighted feature maps and second weighted feature maps to obtain the target tracking result of the current frame. This method can solve the problem that none of the UAV tracking algorithms can well achieve the balance between accuracy and speed.

Tracing and super resolution diffusion are utilized to generate high quality imagery via a trained model. This approach limits the number of tracing iterations to avoid issues with computational complexity. This includes applying a low-resolution image of a first size via a tracing tracing process to a denoise diffusion SR model to generate a resultant high-resolution image of a larger size. The process includes retrieving, at least a portion of imagery data (502) and performing an iterative tracing operation in a plurality of iterations to generate an initial image having a first resolution (504). Then using the first image and a target scaling factor, an iterative super-resolution operation is performed with diffusion and denoising to generate a final image having a second resolution greater than the first resolution, in which the iterative super-resolution operation has multiple iterations that each generate a respective intermediate image having a corresponding noise value (506).

The invention is suitable for the technical field of image processing, and provides a snapshot method, terminal equipment and a computer readable storage medium, and the method comprises the steps: obtaining at least one tracking object; wherein the tracking object is an object tracked according to a shot image of a first camera on the mobile equipment; predicting first snapshot quality of each tracking object in a first future time period; determining a snapshot object from the at least one tracking object according to the first snapshot quality of each tracking object; planning a target snapshot moment for the snapshot object according to the moving speed of the mobile device; controlling a second camera on the mobile device to perform snapshot according to the target snapshot time; wherein the frame rate of the second camera is lower than that of the first camera. By means of the method, the best opportunity of snapshot can be determined, and therefore the snapshot effect is effectively improved.

Embodiments of the present application provide a photoelectric imaging tracing method and apparatus, relating to the technical field of medical devices. The method includes: obtaining an image to be processed; identifying a photoelectric detection head image region in the image to be processed; obtaining an effective region where the photoelectric detection head interacts with the image to be processed in the photoelectric detection head image region; obtaining diagnostic information of the effective region; generating a diagnostic visualization image according to the diagnostic information and the effective region, the diagnostic visualization image including a color region or a contour line; fusing the visualization image and the image to be processed; and displaying the fused image. The present application can improve the problem that when using the photoelectric detection technology for detection, what is seen is not what is detected, and the discrimination accuracy of the region to be detected is not high, and achieve the effect that when using the photoelectric detection technology for detection, what is seen is what is detected, and the discrimination accuracy of the region to be detected is improved.

To provide an information processing device that allows a user to manipulate a virtual object by moving a finger, arm or the like, and is less likely to cause manipulation or display not intended by a user.SOLUTION: An information processing device is provided, comprising image capturing means for acquiring a captured image by capturing an image of real space, tracking means configured to respond to a first case where a first body part of a user is present in the field of view of the image capturing means by tracking the first body part based on the captured image to detect the position and orientation of the first body part, and display control means for controlling the display means to display a first UI in the first case. The first UI is for manipulating a virtual object based on the position and orientation of the first body part detected by the tracking means. When the first body part moves outside the field of view of the image capturing means, the display control means controls display for manipulating the virtual object in accordance with a result of determination as to whether the user intends to manipulate the virtual object.SELECTED DRAWING: Figure 3

The present application belongs to the technical field of X-ray detection, and specifically relates to an image tracking navigation method for an X-ray machine, which can calculate the range of the translation amount according to the geometric information of the tilt angle of the detector and the rotation angle of the circuit board, and then the desired window position. Through precise geometric calculation and image processing algorithm, the present application effectively solves the problem of missing detected points caused by the tilt of the detector or the rotation of the object table during the X-ray detection process. This method can dynamically adjust the detection parameters to ensure that the region of interest always remains within the detection field of view, thereby significantly improving the accuracy and reliability of PCB detection. Especially for small components such as high-density integrated BGA solder joints, this method avoids the detection blind area caused by background interference or changes in viewing angle, making defect identification more accurate and efficient.

The present invention belongs to the technical field of image tracking control, and discloses a data weighted fusion tracking control method for an airborne optoelectronic tracking and aiming system. The method includes measuring the delay of a tracking control loop and establishing a delay model, using the tracking deviation output by a tracker and the angular velocity signal of a fiber optic gyroscope as inputs, and reconstructing a high-bandwidth video tracking control loop. The control commands of the original video tracking control loop and the reconstructed loop are fused and output through data weighting, so as to achieve high-bandwidth control of the video tracking loop of the airborne optoelectronic tracking and aiming system. Compared with the traditional classical PID control, this method introduces more input information, effectively broadens the control bandwidth of the video tracking control loop, and improves the response speed of the optoelectronic tracking and aiming system to target movement.