Abstract: The present invention relates to a method of providing behaviour models of and for dynamic objects. Specifically, the present invention relates to a method and system for generating models and/or control policies for dynamic objects, typically for use in simulators and/or autonomous vehicles. The present invention sets out to provide a set or sets of behaviour models of and for dynamic objects, such as, for example, drivers, pedestrians and cyclists, typically for use in such autonomous vehicle simulators.

Abstract: Devices, systems, and methods for gamifying the process of annotating videos for maintaining engagement in the annotation process and increasing the quality and quantity of an annotated data set. Methods for gamifying annotation of gestures in a video comprise determining gamified feedback based on an input and presenting the gamified feedback to an operator to keep the operator engaged in the annotation process. Methods for gamifying annotation of self-perception gestures in a video by a subject are able to be performed by the subject without the requirement of an operator, in which case the method captures an aspect of the subject's experience and the gamification maintains or increases the subject's engagement with the annotation process. Annotated data sets are used to train artificial intelligence and machine learning systems for automated detection and characterization of subjects' gestures and self-perception gestures.

Abstract: An image signal indicating an inference target image in which a detection target appears is acquired when a domain of the inference target image is different from a domain of a training image or a recognition task of the inference target image is different from a pre-learned task. The image signal is provided to a trained learning model, and from the learning model, an inference time feature amount obtained by combining feature amounts of the detection target in the inference target image after the feature amounts is blurred is acquired. The detection target in the inference target image is recognized on the basis of a representative feature amount that is a registered feature amount of the detection target in an image for conversion in which a domain and a recognition task of the image are the same as those of the inference target image, and the inference time feature amount.

Abstract: Vision Transformers (ViT) have shown their competitive advantages performance-wise compared to convolutional neural networks (CNNs) though they often come with high computational costs. Methods, systems and techniques herein learn instance-dependent attention patterns, utilizing a lightweight connectivity predictor module to estimate a connectivity score of each pair of tokens. Intuitively, two tokens have high connectivity scores if the features are considered relevant either spatially or semantically. As each token only attends to a small number of other tokens, the binarized connectivity masks are often very sparse by nature providing an opportunity to accelerate the network via sparse computations. Equipped with the learned unstructured attention pattern, sparse attention ViT produces a superior Pareto-optimal trade-off between FLOPs and top-1 accuracy on ImageNet compared to token sparsity (48%˜69% FLOPs reduction of MHSA; accuracy drop within 0.4%).

Abstract: One example method includes a machine-learning (ML) model receiving a first input that includes images that have been extracted from a web page and a second input that includes alt-texts that have been extracted from the web page. The alt-texts describe the images. The ML model converts the images into a first embedding representation and converts the alt-texts into a second embedding representation. Based on the first and second embedding representations, a similarity score between the images and the alt-texts is calculated. The similarity score specifies how accurately each of the alt-texts describe the images. The one of the alt-texts having the highest similarity score is then selected.

Abstract: A data analysis system according to this invention includes a data acquirer; an analyzer configured to analyze a to-be-analyzed data by using a learned model; a learned model producer configured to produce the learned model; a storage configured to store each of a plurality of learned models associated with a different version(s) of learned model(s) that is/are other learned model(s) of the plurality of learned models; a display configured to display the learned models; and a controller configured to control the display to display a selected learned model, and the different version(s) of learned model(s) that is/are associated with the selected learned model on a common screen.

Abstract: Encoding a first 3D point cloud of a first coordinate system into a first encoded map comprising first feature center points and first feature vectors; encoding a second 3D point cloud of a second coordinate system into a second encoded map comprising second feature center points and second feature vectors; adapting the first input feature vectors based on the first input feature vectors and the second input feature vectors, to obtain a first joint map; adapting the second input feature vectors based on the first input feature vectors and the second input feature vectors to obtain a second joint map; checking whether a correlation condition is fulfilled; extracting the coordinates of the first joint feature center points and the second joint feature center points if the correlation condition is fulfilled; calculating a transformation between the first coordinate system and the second coordinate system based on the pairs of extracted coordinates.

Abstract: Improved methods for executing multi-layer machine learning model architectures in the context of ordered inputs sequences that experience progressive updates are provided that exhibit increased decreased inference compute cost and/or decreased inference time latency. These improved models include storing some or all of the intermediate outputs of the model's units for later re-use, e.g., once one or more novel inputs of an input sequence have been obtained. Storing such intermediate outputs allows the computational effort used to generate them (e.g., by applying the relevant model input(s) to the relevant unit(s) and/or layer(s) of the model) to be avoided in subsequent execution of the model. Instead, only those model units whose outputs would differ from one model execution to the next are re-computed in order to generate an updated model output, thereby significantly reducing the computational cost and/or time to execute the model in light of the updated input(s).

Abstract: An image identification system includes a first camera that includes a mask and an image sensor, the mask having a changeable mask pattern having a plurality of pinholes, and captures a computational image that is an image with blurring, an image identification unit that identifies the computational image using an identification model that uses the computational image captured by the first camera as input data and an identification result as output data, a mask identification unit that, after the mask pattern is changed, identifies the mask pattern that has been changed, and an identification model change unit that changes the identification model in accordance with the mask pattern identified by the mask identification unit.

Abstract: An image analysis apparatus (102) includes: an icon placement unit (107) accepting an instruction for placing, on a screen, a plurality of icons including a plurality of input source icons each indicating an input source of image data being a target of analysis, a plurality of image processing icons each indicating an image processing engine for the image data, and at least one output destination icon indicating an output destination of a processing result by an image processing engine, and placing the plurality of icons on the screen in accordance with the instruction; a connection unit (110) accepting a connection instruction for connecting icons placed on the screen; and a data flow setting unit (111) setting a flow of data between the icons in accordance with the connection instruction and displaying the flow on the screen.

Abstract: A computer-based system (210) for delineating agricultural fields based on satellite images includes a first subsystem (201) configured to receive at least one multitemporal, multispectral satellite image sequence (101) and pre-processing the images in the at least one multitemporal, multispectral satellite image sequence to generate a pre-processed image sequence (303) of multitemporal multispectral images covering a specific geographical region; a second subsystem (202) configured to perform a super-resolution method on the images in the pre-processed image sequence to generate a high-resolution image sequence (403) of multitemporal multispectral images where corresponding pixel positions in images in the sequence relate to the same geographical ground position; and a third subsystem (203) including a delineating artificial neural network (501) trained to classify pixel positions in the high-resolution image sequence (403) as being associated with a geographical ground position that is part of an agricultur

Abstract: Described herein are various technologies pertaining to an agricultural analysis system for simulating various aspects of the agricultural process. Specifically, an agricultural analysis application is provided that receives target crop growth parameters representative of desired outcomes for a particular crop and a crop growth location indicative of the area to be analyzed by the agricultural analysis application. The agricultural analysis application then identifies satellite image data of the crop growth location and uses the real satellite image data to generate simulated satellite image data using a canopy reflectance simulator. The agricultural analysis application then determines certain parameters that, when provided as input into the canopy reflectance simulator, cause the generated simulated satellite image data to correspond the real satellite image data.

Abstract: Aspects include methods and apparatuses generally relating to agricultural technology and artificial intelligence and, more particularly, to counting and sizing plants in a field.

Abstract: A system for capturing images for training an item identification model obtains an identifier of an item. The system detects a triggering event at a platform, where the triggering event corresponds to a user placing the item on a platform. At least one camera captures an image of the item. The system extracts a set of features associated with the item from the image. The system associates the item to the identifier and the set of features. The system adds a new entry to a training dataset of the item identification model, where the new entry represents the item labeled with the identifier and the set of features.

Abstract: A system for capturing images for training an item identification model obtains an identifier of an item. The system detects a triggering event at a platform, where the triggering event corresponds to a user placing the item on a platform. At least one camera to captures at least one image of the item. The system extracts a set of features associated with the item from the at least one image. The system associates the item to the identifier and the set of features. The system adds a new entry to a training dataset of the item identification model, where the new entry represents the item labeled with the identifier and the set of features.

Abstract: Various embodiments described herein provide for analysis of a video using a scanning technique. According to some embodiments, a video is analyzed by scanning a region of interest in a series of frames of the video, and generating a composite image based on the pixels captured by the scanning operation.

Abstract: A computer vision system, with at least one processor configured to: acquire, from a sports match video, moving image data of a first period and moving image data of a second period; by using a first machine learning model, generate, based on the moving image data of the first period, first estimation data and second estimation data for an estimation period; by using a second machine learning model, generate, based on the moving image data of the second period, the first estimation data and the second estimation data for the estimation period; and generate determination data based on the first estimation data and the second estimation data that are output from the first machine learning model and the first estimation data and the second estimation data that are output from the second machine learning model.

Abstract: Techniques include detecting a person in an area having a target region and annotating a source region surrounding the person according to a hygiene state. When a collision between the source region and the target region is detected, an indicator indicates a violation or a non-violation according to the hygiene state. In some implementations, the hygiene state is one of clean, contaminated, and in process. When the person enters the room, the source region is in a contaminated state until they have performed a hygienic act (e.g., hand washing); after performing the hygienic act, the source region is in a clean state. If a source region in a clean state intersects a target region (e.g., area around a patient's bed), then the indicator indicates a non-violation. If a source region in a contaminated state intersects a target region, then the indicator indicates a violation.

Abstract: A method, information processing device, and an information processing system are provided for automatically and accurately recording information about medical devices in an operating room. A surgical space image is captured by an imaging device, wherein the image includes monitor screens of a plurality of display devices dispersedly located in the surgical space. Measurement information is extracted from the monitor screen of each of the display devices on the basis of the acquired surgical space image. The extracted information regarding the monitor screen of each of the display devices is stored in time series to collectively manage the medical devices.

Abstract: A surveillance information generation apparatus (2000) includes a first surveillance image acquisition unit (2020), a second surveillance image acquisition unit (2040), and a generation unit (2060). The first surveillance image acquisition unit (2020) acquires a first surveillance image (12) generated by a fixed camera (10). The second surveillance image acquisition unit (2040) acquires a second surveillance image (22) generated by a moving camera (20). The generation unit (2060) generates surveillance information (30) relating to object surveillance, using the first surveillance image (12) and first surveillance information (14).

Abstract: Traffic circles are equipped with one or more cameras at its entrances of the traffic circles. The cameras, connected with sensors and central control units, determine speed and moving directions of vehicles entering or circling inside the traffic circle, based on artificial intelligence and machine learning. The speed of a vehicles can be determined by tracking a distance of the car to the camera and a distance between the camera lens and an image sensor inside the camera. The moving direction of the car can be determined by tracking the moving pattern of the car inside the traffic circle, among other ways. The central control determines the traffic pattern inside the traffic circle, gives signals like flashing lights, to vehicles entering the traffic circle to stop and yield to emergency vehicles, pedestrian crossing, or other vehicles approaching the entrance if the incoming vehicles do not have clearances to enter.

Abstract: Described herein are systems, methods, and software to monitor traffic trends associated with objects in a physical area using multiple video sources. In one implementation, a method includes receiving the video from the plurality of video sources and identifying landmark points of the physical area in the video. The landmark points correspond to objects represented in a map. The method also includes monitoring traffic in the physical area from the video and determining a trend in the traffic relative to the landmark points. The method further includes presenting the trend on the map relative to the objects.

Abstract: A refuse vehicle can include a first camera, a second camera, and one or more processing circuits. The one or more processing circuits can communicate with the first camera and the second camera. The one or more processing circuits can receive first data associated with a collection site proximate to the refuse vehicle, detect a first object and a second object based on the first data, generate one or more bounding boxes to establish a dimension for the first object, and determine a position of the first object relative to the second object.

Abstract: 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.

Abstract: A method is provided for predicting respective trajectories of a plurality of road users. Trajectory characteristics of the road users are determined with respect to a host vehicle via a perception system, wherein the trajectory characteristics are provided as a joint vector describing respective dynamics of each of the road users for a predefined number of time steps. The joint vector of the trajectory characteristics is encoded via an algorithm which included an attention algorithm for modelling interactions of the road users. The encoded trajectory characteristics and encoded static environment data obtained for the host vehicle are fused in order to provide fused encoded features. The fused encoded features are decoded in order to predict the respective trajectory of each of the road users for a predetermined number of future time steps.

Abstract: This disclosure relates generally to method and system for multi-object tracking and navigation without pre-sequencing. Multi-object navigation is an embodied Al task where object navigation only searches for an instance of at least one target object where a robot localizes an instance to locate target objects associated with an environment. The method of the present disclosure employs a deep reinforcement learning (DRL) based framework for sequence agnostic multi-object navigation. The robot receives from an actor critic network a deterministic local policy to compute a low-level navigational action to navigate along a shortest path calculated from a current location of the robot to the long-term goal to reach the target object. Here, a deep reinforcement learning network is trained to assign the robot with a computed reward function when the navigational action is performed by the robot to reach an instance of the plurality of target objects.

Abstract: A system for providing situational awareness to a cyclist or other user of a micromobility vehicle comprises a stereo camera assembly and processing logic configured to determine, based on images acquired by the stereo camera assembly, a distance between the cyclist and an object of interest (e.g., a vehicle). The system is configured to determine a threat level of the object based one or more factors such as, e.g., a speed of the object and/or a category of the object. In some examples, the system includes a display and/or an audio indicator to convey information to the cyclist about detected threats. In some examples, the system is configured to produce an audio indication in response to a threat exceeding a threshold threat level. A software platform may be configured to store and/or process micromobility data gathered from one or more users.

Abstract: Computer-implemented method for predicting one or more turn points related to a road a vehicle is travelling on, the one or more turn points indicating locations where the vehicle can change direction, the method comprising: obtaining training images of roads and their environment; receiving labels associated with the roads in the training images, each label comprising a training turn marker; training an artificial neural network on a training dataset to predict one or more turn points, wherein the training dataset comprises the received labels and the obtained training images; recording at least one road image of a road and its environment; and processing the road image by the artificial neural network to predict one or more turn points on the road image.

Abstract: A computer includes a processor and a memory storing instructions executable by the processor to determine an estimated pose of a vehicle in a global reference frame having a first error and to determine an estimated second error based on a combination of the first error and a map error, in which the map error represents a difference between the estimated pose of the vehicle in the global reference frame and a corresponding estimated pose in a map-referenced frame. The stored instructions being additionally to predict a third error in a future map-referenced measurement frame based on the estimated second error and a motion model and to compute an update to the third error by combining the predicted third error with an accumulation of instantaneous vehicle position and heading errors obtained via a comparison between a camera-observed feature and a corresponding feature from a digital map.

Abstract: In various examples, object fence corresponding to objects detected by an ego-vehicle may be used to determine overlap of the object fences with lanes on a driving surface. A lane mask may be generated corresponding to the lanes on the driving surface, and the object fences may be compared to the lanes of the lane mask to determine the overlap. Where an object fence is located in more than one lane, a boundary scoring approach may be used to determine a ratio of overlap of the boundary fence, and thus the object, with each of the lanes. The overlap with one or more lanes for each object may be used to determine lane assignments for the objects, and the lane assignments may be used by the ego-vehicle to determine a path or trajectory along the driving surface.

Abstract: In various examples, object fence corresponding to objects detected by an ego-vehicle may be used to determine overlap of the object fences with lanes on a driving surface. A lane mask may be generated corresponding to the lanes on the driving surface, and the object fences may be compared to the lanes of the lane mask to determine the overlap. Where an object fence is located in more than one lane, a boundary scoring approach may be used to determine a ratio of overlap of the boundary fence, and thus the object, with each of the lanes. The overlap with one or more lanes for each object may be used to determine lane assignments for the objects, and the lane assignments may be used by the ego-vehicle to determine a path or trajectory along the driving surface.

Abstract: There is provided a computer implemented method of monitoring an occupant of a vehicle, comprising: computing a target movement vector set of a skeleton representation from a target video captured by a camera oriented towards a side of the occupant, feeding the target movement vector set of the target skeleton representation into a personalized ML model, obtaining a likelihood of the occupant about to perform an unsafe action as an outcome of the personalized ML model, and generating a feedback by a user interface device in response to the outcome, the feedback generated prior to the occupant performing the unsafe action, thereby preventing the occupant from performing the unsafe action, wherein the personalized ML model is trained on sequences of personalized motions performed by the occupant prior to the performance of the unsafe action and ground truth labels indicating the unsafe action.

Abstract: Systems, methods, and non-transitory computer-readable media for detecting a driver's gaze direction while driving a vehicle are disclosed. At least one processor may be configured to receive image information from an image sensor, detect the vehicle driver in the image information, detect the driver's gaze direction toward a first direction in the image information, predict an amount of time it will take for the driver to shift the gaze direction toward a second direction, using information associated with the detected gaze direction of driver, and generate a message or a command based on the predicted amount of time.

Abstract: Disclosed are a multi-task training technique and resulting model for detecting distracted driving. In one embodiment, a method is disclosed comprising inputting a plurality of labeled examples into a multi-task network, the multi-task network comprising: a backbone network, the backbone network generating one or more feature vectors corresponding to each of the labeled examples, and a plurality of prediction heads coupled to the backbone network; minimizing a joint loss based on outputs of the plurality of prediction heads, the minimizing the joint loss causing a change in parameters of the backbone network; and storing a distraction classification model after minimizing the joint loss, the distraction classification model comprising the parameters of the backbone network and parameters of at least one of the prediction heads.

Abstract: A method, mobile user device and system for identifying authenticity of a cylindrical object from photographic images. The method includes acquiring two or more photographic images of the cylindrical object from different angles around the cylinder axis (A) with an imaging device, generating a target image from the two or more photographic images by image stitching, analysing the target image in relation to a reference image representing an original cylindrical object and generating an identification output based on the analysing, and generating an authenticity identification indication based on the identification output.

Abstract: A part machining feature recognition method based on machine vision learning recognition comprises sample training and part machining feature recognition. The sample training specifically refers to obtaining the 2D images of different 3D models of parts at different angles, marking machining feature information, constructing a 2D image sample, and then performing feature recognition training on the image recognition model using the 2D image samples; the part machining feature recognition specifically refers to taking 2D image screenshots from multiple view angles, recognizing all machining features from 2D image screenshots using the trained image recognition model, mapping the recognized machining features to the 3D model of parts with machining features to be recognized based on the view angle relationship, marking the machining features and geometric surfaces contained in each feature, and completing the automatic recognition of part machining features.

Abstract: In various examples, generating maps using first sensor data and then annotating second sensor data using the maps for autonomous systems and applications is described herein. Systems and methods are disclosed that automatically propagate annotations associated with the first sensor data generated using a first type of sensor, such as a LiDAR sensor, to the second sensor data generated using a second type of sensor, such as an image sensor(s). To propagate the annotations, the first type of sensor data may be used to generate a map, where the map represents the locations of static objects as well as the locations of dynamic objects at various instances in time. The map and annotations associated with the first sensor data may then be used to annotate the second sensor data and/or determine additional information associated with the objects represented by the second sensors data.

Abstract: An anomaly labeled-assistant detection system and a method thereof are provided. The anomaly labeled-assistant detection system includes a computing apparatus and a storage apparatus. The computing apparatus includes an anomaly labeled detection model, where the computing apparatus detects a plurality of pieces of labeled image data with a labeled category through the anomaly labeled detection model, the anomaly labeled detection model respectively generates an inference category corresponding to each piece of labeled image data, and the computing apparatus compares the labeled category and the inference category according to each piece of labeled image data, and automatically lists the labeled image data as anomaly labeled data when the labeled category of the labeled image data is different from the inference category. The storage apparatus is electrically connected to the computing apparatus, to store the labeled image data.

Abstract: A text recognition system causes a trained region encoder to determine a region of interest of an image file. The system modifies a first image associated with the first region of interest (e.g., parsed out from the first region) to generate a data augmentation entity that includes a modified image. Using a trained instance encoder, the system generates a first set of visual instances corresponding to the first region of interest image and a second set of visual instances corresponding to the data augmentation entity. The system generates the corresponding first and second sequences. By executing a self-supervised contrastive loss function on the first and second sequences, the system automatically updates a continual knowledge distillation model of the trained region encoder. The system provides the first sequence to an instance decoder to generate output text in response to the prompt.

Abstract: An image processing system, comprising at least one processor configured to: detect a plurality of character string regions each including any character string from a target object image relating to a target object including a standard character string and a non-standard character string; determine whether a standard region including the standard character string exists in the plurality of character string regions; and identify, when the standard region is determined to exist, a character string region that has a predetermined positional relationship with the standard region as a non-standard region that includes the non-standard character string.

Abstract: A method and system of extracting one or more non-semantic entities in a document image including data entities is disclosed. The methodology includes extraction, by a processor, of row entities and corresponding row location based on a text extraction technique from the document image. The row entities are split into split-row entities based on a splitting rule. Semantic entities are determined from alphabetic entities using semantic recognition technique. The non-semantic entities are determined as split-row entities other than semantic entities. Feature values of each feature type for each of the non-semantic entities is determined. The processor further determines a first probability output for non-semantic entities and a second probability output for semantic entities surrounding the non-semantic entities. The system further labels each of the non-semantic entities based on determination of a highest probability value from a sum of the first probability output and the second probability output.

Abstract: A method includes receiving, from a user device associated with a user, a plurality of annotated documents. Each respective annotated document includes one or more fields and each respective field labeled by a respective annotation. The method includes, for a threshold number of iterations, randomly selecting a respective subset of annotated documents from the plurality of annotated documents; training a respective model on the respective subset of annotated documents; and generating, using the plurality of annotated documents not selected for the respective subset of annotated documents, a respective evaluation of the respective model. The method also includes providing, to the user device, each respective evaluation.

Abstract: A corrective noise system receives an electronic version of a fillable form generated by a segmentation network and receives a correction to a segmentation error in the electronic version of the fillable form. The corrective noise system is trained to generate noise that represents the correction and superimpose the noise on the fillable form. The corrective noise system is further trained to identify regions in a corpus of forms that are semantically similar to a region that was subject to the correction. The generated noise is propagated to the semantically similar regions in the corpus of forms and the noisy corpus of forms is provided as input to the segmentation network. The noise causes the segmentation network to accurately identify fillable regions in the corpus of forms and output a segmented version of the corpus of forms having improved fidelity without retraining or otherwise modifying the segmentation network.

Abstract: A method for providing non-visual access to graphical artifacts available in digital content includes classifying a graphical artifact into known and/or unknown categories using a deep neural network. The method further includes identifying semantically connected visual and textual components of the graphical artifact, using a deep learning-based object detection model. Furthermore, the method includes extracting the visual and the textual components in a unified framework with predefined semantics associated with each component, using a pre-trained large multi-modal model fine-tuned to extract both the visual and the textual components from an image in the graphical artifact. The method further includes filtering out the predefined semantics through extraction and converting the predefined semantics into accessible representations. Also, the method includes delivering the accessible representations in conformance with requirements of a delivery system.

Abstract: The subject technology provides for stroke based control of handwriting input. The disclosed stroke based control facilitates selection, copy, paste, search, data detection and other operations for handwritten electronic text. The selection of text represented by handwritten strokes can be performed without drawing a lasso or other loop around the desired text, by using known boundaries of words and phrases in stroke space. Selection of text in this manner allows copy and/or paste of recognized words or phrases, of images of the words or phrases, and/or of the strokes themselves. Boundaries, in stroke space, of actionable data represented by the strokes can also allow action options to be provided when a user interacts with strokes within the boundary.

Abstract: Artificial intelligence (AI) systems of the inventive subject matter are directed to receiving uploaded images containing one or more document, identifying the type of documents received, and returning information contained in those documents. Upon receiving an image containing a document, the AI system: checks for barcodes, OCRs any text, detects visual features, and detects an overall document shape. The AI system can then use any information gathered during those steps to ultimately verify information contained in the document.

Abstract: The present disclosure describes image analysis techniques that identify the source of a document. Once the source of the document is determined, the image analysis may locate one or more anchor fields in the document. The anchor fields may identify one or more additional fields that contain time-sensitive data and/or information. The image analysis performed herein may identify the time-sensitive data and/or information and process the data and/or information to schedule due dates and reminders.

Abstract: A system for monitoring occupancy of a structure includes at least one camera configured to acquire an image of the structure; processor circuitry coupled with a memory configured to process the image; and an output device; wherein, in order to process the image, the processor circuitry is configured to: detect an object in the image; extract one or more features of the object; identify whether the object is a person; associate a unique tracker with each identified person; keep track of the identified people in the structure; calculate a load on the structure based on an attribute of each person; and instruct the output device to generate an output signal when the calculated load reaches a predetermined threshold.

Abstract: An image processing apparatus for detecting a person region representing a person in an image comprises one or more memories storing instructions; and one or more processors executing the instructions to obtain an image, generate a plurality of rotated images by rotating an obtained image by a plurality of preset angles, execute person detection processing on each one of the plurality of rotated images and obtain detection information including information indicating a candidate region of a detected person and information indicating a likelihood as a person in the candidate region, and using detection information obtained via detection of the detected person, determine a candidate region to be removed from a person region in accordance with each likelihood and remove the candidate region.

Abstract: This application provides a palm contour extraction method performed by a computer device. The method includes: obtaining bone point information respectively corresponding to palm bone points in a target palm image, the bone point information including bone point positions and bone point types; matching the palm bone points based on the bone point positions and the bone point types, to obtain a plurality of palm bone point sets, each palm bone point set having a corresponding geometry processing type for a corresponding set of palm bone points in the target palm image; determining a plurality of palm contour points from the plurality of palm bone point sets based on their respective geometry processing type; and generating a palm contour corresponding to the target palm image based on the plurality of palm contour points.