Abstract: A system for generating a dark current residual image is configurable to generate a weighted average image by: (i) determining a region-based weight value for each pixel of an input image based upon a light level of a region in which the pixel lies; and (ii) combining the input image with a previous image using the region-based weight values for each pixel of the input image. The system is also configurable to generate a dark current residual image based upon the weighted average image.

Abstract: An underwater image enhancement method and image processing system using the same are provided. The method includes the following steps. An original underwater image is received. An original histogram of the original underwater image is generated. The original histogram is input into a deep learning model to generate an optimized histogram. An optimized underwater image is generated according to the optimized histogram and the original underwater image.

Abstract: An image processing device for a light scanner system for scanning an object, including circuitry configured to: obtain first image data representing a first image of the object that is illuminated with a line of light in an illumination plane, wherein the first image is acquired at a first viewpoint; obtain second image data representing a second image of the illuminated object, wherein the second image is acquired at a second viewpoint being different from the first viewpoint; and project the first and the second image data in the illumination plane representing a projected first and second image, respectively, for merging the first and second image data in the illumination plane.

Abstract: An information processing apparatus that makes it possible to see information in an easily understandable manner when this information is superimposed onto a video image has at least one memory and at least one processor. The at least one memory and the at least one processor are configured to extract a region of a human in an image; superimpose predetermined superimposition information onto the image; determine whether or not the human is performing a predetermined action, and to output determination results; extract a region of the predetermined information that overlaps with the region of the human as an overlapping region based on the determination results, the region of the human, and the predetermined information; and change a transparency of at least a portion of the predetermined information such that the transparency increases according to the overlapping region.

Abstract: Provided are a method and apparatus for generating an image reconstruction model and an image reconstruction method and apparatus using the image reconstruction model. The image reconstruction apparatus extracts, through an encoder configured to extract a structure and a texture, a structure and a texture of each of a first medical image and a second medical image and generates, through a decoder configured to reconstruct an image based on a structure and a texture, a third medical image in which the texture of the second medical image is combined with the structure of the first medical image.

Abstract: Systems and methods for motion artifact simulation are provided. The systems may obtain a target image including a target object. The systems may determine a plurality of sub-periods of a time period corresponding to the target image. The systems may determine a plurality of motion vector fields of the target object in the plurality of sub-periods. Each motion vector field of the plurality of motion vector fields may correspond to one of the plurality of sub-periods. The systems may determine a plurality of reconstruction images of the target object corresponding to the plurality of sub-periods based on projection data of the target image. Each reconstruction image of the plurality of reconstruction images may correspond to one of the plurality of sub-periods. The systems may generate a motion artifact simulation image of the target object based on the plurality of motion vector fields and the plurality of reconstruction images.

Abstract: A system and method for denoising a grey scale image of a pattern on a substrate, including: scanning an electron beam along a pattern, detecting a scan line waveform from the scanned electron beam, obtaining a model scan line waveform from the detected scan line waveform, augmenting the model scan line waveform and adding noise to the augmented model scan line waveform, inputting the noisy augmented model scan line through a deep learning neural network (DNN) process, comparing the output of the DNN process to the augmented model scan line waveform before noise is added, and backpropagating the compared results into the DNN process to obtain a noise discrimination function.

Abstract: Augmented Denoising Diffusion Implicit Models (“DDIMs”) using a latent trajectory optimization process can be used for image generation and manipulation using text input and one or more source images to create an output image. Noise bias and textual bias inherent in the model representing the image and text input is corrected by correcting trajectories previously determined by the model at each step of a diffusion inversion process by iterating multiple starts the trajectories to find determine augmented trajectories that minimizes loss at each step. The trajectories can be used to determine an augmented noise vector, enabling use of an augmented DDIM and resulting in more accurate, stable, and responsive text-based image manipulation.

Abstract: The present disclosure relates to systems, non-transitory computer-readable media, and methods for inpainting digital images utilizing mask-robust machine-learning models. In particular, in one or more embodiments, the disclosed systems obtain an initial mask for an object depicted in a digital image. Additionally, in some embodiments, the disclosed systems generate, utilizing a mask-robust inpainting machine-learning model, an inpainted image from the digital image and the initial mask. Moreover, in some implementations, the disclosed systems generate a relaxed mask that expands the initial mask. Furthermore, in some embodiments, the disclosed systems generate a modified image by compositing the inpainted image and the digital image utilizing the relaxed mask.

Abstract: The present disclosure relates to systems, methods, and non-transitory computer-readable media that generate and implement semantic histories for digital image editing. For instance, in some embodiments, the disclosed systems determine, for a digital image, a first semantic state and a second semantic state that reflects a first semantic change in the digital image from the first semantic state. Additionally, the disclosed systems determine, for the digital image, a third semantic state that reflects a second semantic change in the digital image from the first semantic state. The disclosed systems further generate a semantic history log for the digital image, the semantic history log including a first editing branch corresponding to the second semantic state and a second editing branch corresponding the third semantic state. Based on one or more user interactions with the semantic history log, the disclosed systems modify the digital image.

Abstract: An image enhancement method and an electronic device are disclosed. The image enhancement method includes the following steps. An original image is divided into multiple image blocks. A statistical data of each of the image blocks is calculated, and at least one index is obtained according to the statistical data of the each of the image blocks. A final tone mapping curve of the each of the image blocks is obtained based on multiple predefined tone mapping curves according to the at least one index of the each of the image blocks. Afterwards, the original image is adjusted according to the final tone mapping curve of the each of the image blocks to obtain an enhanced image.

Abstract: Embodiments of this application provide an image processing method and an electronic device, and relate to the field of image processing technologies. The method includes: obtaining display information of a to-be-displayed image at current target backlight luminance of a display and display information of the to-be-displayed image at luminance that is of the display and that is higher than the current target backlight luminance, and obtaining display information of the to-be-displayed image based on the two types of display information.

Abstract: A display device according to an embodiment may include a display and a controller configured to apply a global contrast curve that adjusts a luminance of an entire region of an input image, generate a plurality of local contrast curves corresponding to each of a plurality of local regions based on a local information of each of the plurality of local regions included in the entire region, apply each local contrast curve to each local area, and adjust one or more of a noise processing intensity or a sharpness intensity of a local output image to which the local contrast curve is applied based on each local information.

Abstract: A surface defect detection method for a primary cable of an aerostat based on few-shot learning includes the following steps. A hardware and software system environment is set. A surface image of the primary cable is acquired and processed to obtain augmented surface image data, which is labeled to construct a surface defect sample library. A defect detection network model is designed and constructed, and then is trained based on the surface defect sample library. A query set in the surface defect sample library is processed with the trained defect detection network model to obtain shallow texture features and high-level semantic features. The shallow texture features are transferred to the high-level semantic features through skip connection. The surface defect detection data under different detection operation modes are obtained at a terminal. This application also provides a surface defect detection system.

Abstract: Utilizing pose-dependent machine learning (ML) models in determining whether an anomaly is present within an environment. A first ML model is used to generate first output based on processing a first image, captured by a camera of a mobile robot, when the camera is at a first pose. The first ML model is used based on the first image being captured at the first pose and the first ML model being trained for the first pose. A second ML model is used to generate second output based on processing a second image that is captured, by the camera, when the camera is at a distinct second pose. The second ML model is used based on the second image being captured at the second pose and the second ML model being trained for the second pose. The first and second outputs can be used in determining whether an anomaly is present.

Abstract: A method of generating synthetic data for composite manufacturing inspection comprising creating a three-dimensional model of a composite layup; generating an inconsistency in the three-dimensional model of the composite layup; setting a camera position and a camera orientation in the three-dimensional model space; generating a captured image of the three-dimensional model and the inconsistency from the camera position and the camera orientation; and creating a secondary image from the captured image, the secondary image having colorations of a preselected inspection type.

Abstract: Methods and systems for generating information for use in setting up a process performed on a specimen are provided. One method includes clustering dies on a specimen based on colors assigned to the dies responsive to predicted defect densities in the dies determined from measurements performed on the specimen thereby generating initial die clusters. The method also includes analyzing the initial die clusters in location space to determine if any of the initial die clusters contain two or more die clusters. In addition, the method includes designating the initial die clusters that do not contain two or more die clusters and the two or more die clusters contained in any of the initial die clusters as the final die clusters. The method further includes storing information for the final die clusters for use in setting up a process performed on the specimen.

Abstract: An inspection system for inspecting a display device including pixels includes a camera for capturing the display device and providing a sensing image signal and an inspection device for receiving the sensing image signal and outputting a compensation signal for compensating for luminance of a defective pixel. The inspection device includes an image detector outputting a sensing input signal corresponding to the sensing image signal, a defect coordinate detector detecting the defective pixel based on the sensing input signal and outputting defect coordinates indicating a position of the defective pixel, an image analyzer analyzing consistency of the defect coordinates based on the sensing image signal and the sensing input signal and outputting final coordinates, and a defect compensation calculator outputting a compensation signal for adjusting luminance of compensation pixels adjacent to the defective pixel from among the pixels based on the final coordinates.

Abstract: Implementations process an image, using a machine learning (ML) model, to generate a reduced dimensionality ML output. The image captures component(s) of an environment (e.g., an industrial automation facility) and can be captured via a vision component of a mobile robot within the environment. The ML output indicates, for each of N regions of the image, a corresponding anomaly detection probability that indicates whether an anomaly is present in a respective region, of the N regions of the image. Those implementations determine, based on the ML output, a quantity of anomaly detection probabilities that each satisfy a threshold. Those implementations further determine whether the quantity is both greater than a lower bound value and less than an upper bound value and, if so, cause remediating action(s) to be performed, such as causing rendering of an alert that indicates an anomaly is present for the component(s).

Abstract: According to an image forming apparatus, it is possible to sort a defective product without requiring time and effort of a user and without increasing or decreasing cost. The image forming apparatus includes: a sheet conveyer having a cylindrical shape that conveys a sheet; an image former that forms an image on the sheet conveyed by the sheet conveyer; an image reader that generates a read image of a printed product which is the sheet on which the image has been formed by the image former; and a hardware processor, in which the hardware processor inspects for presence or absence of a defect in the printed product based on the read image, and causes the printed product being inspected to rotate on an outer peripheral portion of the sheet conveyer without ejecting the printed product being inspected from the apparatus during start to end of inspection of the printed product.

Abstract: Provided are an information processing apparatus, an information processing method, and an information processing program with which a state of a manufactured component can be accurately understood. An information processing apparatus includes a processor, in which the processor acquires first information including information on severity of a discontinuity obtained by analyzing an image obtained by imaging an object, calculates a distribution of the severity of the discontinuity based on the first information, acquires information on a threshold value of a defect, generates second information in which the distribution of the severity of the discontinuity and the threshold value are displayed, and displays the second information on a display device.

Abstract: An immersive scene monitoring system presents a monitoring interface providing a virtual reality (VR)/augmented reality (AR) view of a scene associated with production operation(s) for a (e.g., pharmaceutical) product. The VR/AR view of the scene may be displayed with interactable graphical elements representing input fields corresponding to attributes of quality management events to be generated. The system generates the quality management events based on input (e.g., voice, gesture) received via the monitoring interface. The VR and/or AR view of the scene may be displayed with indicators of detected anomalies, and the input fields and/or corresponding attributes may be automatically populated by the system based on recognized objects and/or anomalies depicted in captured image data of the VR/AR view of the scene. The system can be configured for different scenes pertaining to different production operations and/or different products via a configuration interface.

Abstract: Provided is a visual inspection device including: an image pickup unit configured to pick up an image of a surface of an object to be inspected; a posture control unit configured to change a posture of the object to be inspected with respect to the image pickup unit; a storage unit having stored therein a learning result obtained by machine learning using a defective product sample image, the defective product sample image being generated by combining a defect portion image prepared in advance with a non-defective product image of the object to be inspected at a combination position obtained by referring to three-dimensional data corresponding to the object to be inspected tilted so as to form a posture angle set in advance; and an inspection unit configured to inspect the surface of the object to be inspected based on a picked-up image picked up at the posture angle by the image pickup unit and on the learning result.

Abstract: A method for defect probability estimation based on relationships with concepts, the method may include (a) obtaining an evaluated patch representation, and (b) determining that the evaluates patch representation is not faulty when at least one of the following occurs: (a) for each RPR of a first number (N1) of RPRs, a similarity between the evaluated patch representation is not lower than the first RPR similarity threshold of the RPR; or (b) for each RPR of a second number (N2) of RPRs, a similarity between the evaluated patch representation is lower than the first RPR similarity threshold and not lower than the second RPR similarity threshold of the RPR.

Abstract: Disclosed is an intra image processing method including obtaining a first intraoral image, and obtaining a second intraoral image, into which input additional information is embedded, by embedding the input additional information into at least a partial area of the first intraoral image.

Abstract: A medical image processing apparatus comprising processing circuitry configured to: receive computed tomography (CT) perfusion image data that is representative of an anatomical region of a patient or other subject, determine or define a region of interest of the anatomical region, in-paint one or more parts of the anatomical region, the one or more parts being located inside and/or outside of the region of interest and perform low-pass filtering of at least the region of interest to generate filtered CT perfusion image data.

Abstract: Medical image data is received at the system from a plurality of public or private datasets; An AI model is trained on the datasets to learn image classification and outputs (i) an image-level classification function, (ii) an object-level classification function, and (iii) a plurality of image classification weights; the AI model is trained on the datasets to learn image localization and output an object localization function and image localization weights; the AI model is trained on the datasets to learn image segmentation and output an object segmentation function and image segmentation weights; each of the image classification weights is integrated with the image localization weights and the image segmentation weights into a single pre-trained AI model; each of the image-level classification function, the object-level classification function, the object localization function and the object segmentation function are integrated into a single pre-trained AI model for use with medical image analysis.

Abstract: A radiographic image analysis apparatus includes a hardware processor. The hardware processor estimates a respiratory function of a subject based on a feature amount of a region of interest extracted from a dynamic image of a chest of the subject and examination result information on a past respiratory function examination on the subject.

Abstract: A surgery assisting system includes an image acquisition unit capturing endoscope images, a surgical technique selection reception unit for selecting a surgical technique, an organ area estimation unit utilizing an organ area estimation model to predict the position and range of a potentially damaged organ based on the selected surgical technique and captured image, and an estimated area display unit overlaying the predicted organ damage information onto the captured image. This system enhances surgical planning by providing real-time visualization of potential organ damage areas, aiding surgeons in performing procedures with increased precision and safety.

Abstract: A rapid and automatic virus imaging and analysis system includes (i) electron optical sub-systems (EOSs), each of which has a large field of view (FOV) and is capable of instant magnification switching for rapidly scanning a virus sample; (ii) sample management sub-systems (SMSs), each of which automatically loads virus samples into one of the EOSs for virus sample scanning and then unloads the virus samples from the EOS after the virus sample scanning is completed; (iii) virus detection and classification sub-systems (VDCSs), each of which automatically detects and classifies a virus based on images from the EOS virus sample scanning; and (iv) a cloud-based collaboration sub-system for analyzing the virus sample scanning images, storing images from the EOS virus sample scanning, and storing and analyzing machine data associated with the EOSs, the SMSs, and the VDCSs.

Abstract: There is provided an image processing apparatus including: at least one processor, in which the processor is configured to execute: calcification candidate image detection processing of detecting a calcification candidate image estimated to be a calcification image from a series of a plurality of projection images obtained by tomosynthesis imaging of a breast or a plurality of tomographic images obtained from the plurality of projection images; calcification candidate image group generation processing of generating a calcification candidate image group by cutting out a region including the calcification candidate image detected by the calcification candidate image detection processing, from each of the plurality of projection images; and calcification determination processing of determining whether or not the calcification candidate image is a calcification image based on the calcification candidate image group generated by the calcification candidate image group generation processing.

Abstract: There is provided an image processing apparatus that executes calcification image detection processing of detecting a calcification image based on a plurality of tomographic images obtained from a series of a plurality of projection images; region-of-interest image generation processing of generating a region-of-interest image based on a detection result by the calcification image detection processing by cutting out a region including the calcification image, from a projection image obtained at an irradiation position among the plurality of projection images or the plurality of tomographic images, the irradiation position being closest to a position facing a detection surface of a radiation detector; shape restoration processing of restoring a shape of the calcification image based on the region-of-interest image generated by the region-of-interest image generation processing; and synthesized two-dimensional image generation processing of generating a synthesized two-dimensional image based on a shape restora

Abstract: Techniques are for detecting presence of a problematic cellular entity in a target. In an example, using an analysis model, a fluorescence-based image is analyzed. The analysis model is trained using a number of reference fluorescence-based images for detecting the presence of problematic cellular entities in targets. Based on the analysis, a problematic cellular entity present in the target is detected. To perform the detection, the analysis model is trained to differentiate between the fluorescence in the fluorescence-based image emerging from the problematic cellular entity and the fluorescence in the fluorescence-based image emerging from regions other than the problematic cellular entity.

Abstract: According to an embodiment, a medical image processing device includes a first image acquirer, a second image acquirer, a treatment error acquirer, a difference calculator, and a differential statistical quantity calculator. The first image acquirer acquires a first fluoroscopic image of a patient. The second image acquirer acquires a second fluoroscopic image photographed at a timing different from the first fluoroscopic image. The treatment error acquirer acquires a treatment error occurring when an alignment process is performed or a treatment error occurring in treatment. The difference calculator calculates a difference image between the second fluoroscopic image to which the virtual perturbation is applied at the position of the patient shown therein based on the treatment error and the first fluoroscopic image.

Abstract: A medical image processing device according to an embodiment has a first image acquirer, a second image acquirer, a 3D-3D positioning executer, and a display controller. The first image acquirer acquires a first three-dimensional fluoroscopic image that is a three-dimensional fluoroscopic image of a patient. The second image acquirer acquires a second three-dimensional fluoroscopic image that is a three-dimensional fluoroscopic image of the patient. The 3D-3D positioning executer calculates a first shift amount between the first three-dimensional fluoroscopic image and the second three-dimensional fluoroscopic image. The display controller causes the display device to display a first DRR image generated from the second three-dimensional fluoroscopic image corrected based on the first shift amount, and a two-dimensional fluoroscopic image of the patient.

Abstract: The present disclosure relates to a deep learning neural network that can identify corpora lutea in the ovaries and a rules-based technique that can count the corpora lutea identified in the ovaries and infer an ovarian toxicity of a compound based on the count of the corpora lutea (CL). Particularly, aspects of the present disclosure are directed to obtaining a set of images of tissue slices from ovaries treated with an amount of a compound; generating, using a neural network model, the set of images with a bounding box around objects that are identified as the CL within the set of images based on coordinates predicted for the bounding box; counting the bounding boxes within the set of images to obtain a CL count for the ovaries; and determining an ovarian toxicity of the compound at the amount based on the CL count.

Abstract: The present invention provides a method for detecting tissue hemorrhage with image analysis. A host produces a plurality of hyperspectral image information according to a plurality of reference images. An image extraction unit extracts an input image to the host. The host transforms the input image according to the plurality of hyperspectral image information to produce a hyperspectral input image. The host produces an input image spectrum according to the hyperspectral input image. The host performs a feature operation on the input image spectrum according to a preset cell band corresponding to a surface cell of small intestine for generating a plurality of corresponding feature bands. The host performs at least one convolution operation on the plurality of feature bands according to a plurality of kernels for producing a convolution result.

Abstract: A display control apparatus includes an inputter for allowing a user to input a query image to initiate a diagnostic process, a processor configured to input the query image into a disease identifier for generating a plurality of indexes, acquire a malignant index of the plurality of indexes representing a possibility that an attribute of a disease of a diagnosis target area is malignant and a first disease attribute index of the plurality of indexes representing a possibility that an attribute of the disease of the diagnosis target area is a prescribed first disease attribute, and cause the acquired malignant index and the acquired first disease attribute index to be displayed in association with each other on a display.

Abstract: A tumor classification method using multi-wavelength photoacoustic image and ultrasound image comprises the steps of: receiving, by an analysis device, an image set including photoacoustic frames and ultrasound frames collected over time for a predetermined period of time by each of a plurality of wavelength bands with respect to a subject; selecting, by the analysis device, photoacoustic frames and ultrasound frames to be analyzed, from the image set, on the basis of ultrasound image; performing, by the analysis device, spectral unmixing analysis on the selected photoacoustic frames; calculating, by the analysis device, at least one parameter for a tumor region by using the result of the spectral unmixing analysis; and classifying, by the analysis device, the subject's tumor by using the parameter.

Abstract: A system for 3D medical image segmentation includes a medical imaging device for obtaining a plurality of 2D images forming a volumetric image, processing circuitry, and a display. The processing circuitry is configured with a first stage to divide the volumetric image into 3D image patches, a hierarchical encoder-decoder structure in which resolution of features of the 3D image patches is decreased by a factor of two in each of a plurality of stages of the encoder, an encoder output connected to the decoder via skip connections, and a convolutional block to produce a voxel-wise final segmentation mask. The encoder includes a plurality of efficient paired attention blocks each with a spatial attention branch and a channel attention branch that learn respective spatial and channel attention feature maps. The display displays the final segmentation mask.

Abstract: A method for acquiring one or multiple two-dimensional magnetic resonance image of a slice through a region of interest may include receiving a gradient field map defining a transformation from a real space to a gradient space, the gradient space being distorted with respect to the real space due to a non-linearity of the gradients; receiving a nominal size, position and orientation of a target slice in real space; transforming a set of points within the target slice into gradient space, resulting in a distorted set of points defining a distorted slice in gradient space; calculating the position and orientation of a new slice, which approximates the position and orientation of the distorted slice, by shifting and/or tilting the target slice; acquiring a two-dimensional magnetic resonance image by excitation of the new slice; and performing a two-dimensional distortion correction of the two-dimensional image to remove in-plane distortions.

Abstract: The present disclosure describes systems, non-transitory computer-readable media, and methods for detecting and indicating modifications between a digital image and a modified version of a digital image. For example, the disclosed systems generates an ordered collection of change records in response to detecting modifications to the digital image. The disclosed systems generates determine one or more non-contiguous modified regions of pixels in the digital image based on the change records. The disclosed system generate an edited region indicator corresponding to the non-contiguous modified regions. The disclosed systems can further color-code the edited region indicator at an object level based on objects in the modified version of the digital image.

Abstract: The present disclosure relates to systems, non-transitory computer-readable media, and methods for utilizing machine learning to generate a mask for an object portrayed in a digital image. For example, in some embodiments, the disclosed systems utilize a neural network to generate an image feature representation from the digital image. The disclosed systems can receive a selection input identifying one or more pixels corresponding to the object. In addition, in some implementations, the disclosed systems generate a modified feature representation by integrating the selection input into the image feature representation. Moreover, in one or more embodiments, the disclosed systems utilize an additional neural network to generate a plurality of masking proposals for the object from the modified feature representation. Furthermore, in some embodiments, the disclosed systems utilize a further neural network to generate the mask for the object from the modified feature representation and/or the masking proposals.

Abstract: Systems and methods of the inventive subject matter are directed to artificial intelligence systems that are configured to receive an uploaded file from a user, where that upload contains images of one or more documents. Documents can be, e.g., receipts, identification cards, and so on. Once received, the system converts the image to a tensor to facilitate identifying mask images. The AI system then converts any identified mask images to grayscale before converting the image to black and white. White spaces are identified, and contours are identified that surround the white spaces. Next, white spaces with areas below a threshold size are discarded and white spaces with areas above that threshold have their contours stored. The AI system finds a minimum bounding rectangle for each contour and then crops the image around each white space accordingly. In this way, documents can be found within an image file and segmented to later processing.

Abstract: A computer that includes a processor and a memory, the memory including instructions executable by the processor to generate background pixels and object pixels and generate background pixel ray data based on the background pixels and object ray pixel data based on the object pixels. The background pixel ray data can be input to a first neural network to generate background neural radiance fields (NeRFs) and the object pixel ray data can be input to a second neural network to generate object NeRFs. An output image can be rendered based on the background NeRFs and the object NeRFs.

Abstract: A system configured to obtain, by a processor, a data stream from a camera of a client device and identifying an object in the data stream. One or more operations are performed to crop the object from background data of the data stream. On a user interface of the client device, a cropped version of the object is then presented.

Abstract: A method of an optical sensor includes: using a single image sensor array of the optical sensor to capture and generate an image frame; selecting a plurality of image regions within the image frame, the plurality of image regions being not overlapped; detecting a plurality of motion results of the plurality of image regions respectively and individually based on images of the plurality of image regions without performing object detection; and determining a plurality of motions of surface feature images respectively within the plurality of image regions according to the plurality of motion results of the plurality of image regions.

Abstract: An example non-transitory machine-readable medium includes instructions that, when executed by a processor, cause the processor to analyze a video captured by a computing device to detect a sequence of motion in the video, and select an audio endpoint of the computing device based on the sequence of motion.

Abstract: A system and associated methods are disclosed for automatically detecting, tracking and classifying an at least one object, moving within an environment proximal to a structure, using a camera. In at least one embodiment, a processor receives a plurality of video frames as captured by the camera. For each of the video frames, the processor detects the presence of the at least one object within said video frame, creates a track for each of the at least one detected object, identifies each track as either an acceptable track or noise, classifies each of the acceptable tracks, and generates an at least one summary image that visually represents the at least one detected object within the environment proximal to the structure, allowing for validation and performance evaluation of the object detection, tracking, and classification process.

Abstract: In some examples, systems and methods for multiple-sensor object tracking are provided. For example, a method includes: receiving a first sensor feed and a second sensor feed from a plurality of sensors respectively. The first sensor feed includes a set of first images. The second sensor feed includes a set of second images. In some examples, the method further includes generating an image transformation based on at least one first image in the set of first images and at least one second image in the set of second images, applying the image transformation to the set of second images, aggregating the set of first images and the set of transformed second images to generate a set of aggregated images, and applying a multiple object tracking model to the set of aggregated images to identify a plurality of objects.