Exploiting hyperbolic distance learning of hierarchical structures to improve open-world object detection.

Abstract

An open-world object detection procedure includes receiving object data containing embedding data corresponding to embeddings for known objects in a first input image, projecting the embeddings into a hyperbolic embedding space containing embeddings in a plurality of categories of objects, each containing one or more classes of objects, regularizing the projected embeddings in the hyperbolic embedding space by moving each of the projected embeddings closer to embeddings in the same category of the plurality of categories and farther away from embeddings in different categories of the plurality of categories, receiving an unmatched query corresponding to an object in a second input image, and generating, based on the hyperbolic embedding space containing the regularized embeddings, an output signal indicatingwhether the object in the second input image corresponds to an unknown object in one of the classes of objects.

Description

Technical area

[0001] The present disclosure relates to open world object detection (OWOD) and, more particularly, to systems and methods for operating devices implementing OWOD. background

[0002] Advances in object detection (OD) have unlocked a wealth of practical applications, such as robotics and self-driving cars. Breakthroughs in attention-based neural network architectures, such as deformable transformers, have yielded impressive performance in these situations. However, many of these solutions assume a fixed number of classes (a closed-world assumption), which is rare in reality. Continuous object detection incrementally adds new classes, resulting in a distributional shift in the input and the well-known phenomenon of "catastrophic forgetting," which refers to the network forgetting previously learned knowledge.

[0003] Open-world object detection (OWOD) is a challenging and realistic task that goes beyond the scope of standard object detection tasks. Open-world systems introduce the detection and integration of new classes. Accordingly, OWOD involves detecting both known and unknown objects while integrating learned knowledge for future tasks. Brief description

[0004] A method for performing open-world object detection includes receiving object data containing embedding data corresponding to embeddings for known objects in a first input image, projecting the embeddings into a hyperbolic embedding space containing embeddings in a plurality of categories of objects, each containing one or more classes of objects, regularizing the projected embeddings in the hyperbolic embedding space by moving each of the projected embeddings closer to embeddings in the same category of the plurality of categories and farther from embeddings in different categories of the plurality of categories, receiving an unmatched query corresponding to an object in a second input image, and generating, based on the hyperbolic embedding space containing the regularized embeddings, an output signal indicatingwhether the object in the second input image corresponds to an unknown object in one of the classes of objects.,

[0005] According to further features, the object data includes a set of bounding boxes and class labels for the known objects in the first input image. The first input image corresponds to a training image. The method further comprises determining a hyperbolic contrastive loss corresponding to the hyperbolic embedding space, and regularizing the projected embeddings comprises regularizing the projected embeddings based on the hyperbolic contrastive loss. Regularizing the projected embeddings comprises determining respective hyperbolic averages of embeddings in each of the classes of objects. Generating the output signal comprises determining a threshold distance based on the hyperbolic averages and generating the output signal based in part on the threshold distance.Generating the output signal includes determining whether the unmatched sample is less than the threshold distance from at least one of the hyperbolic averages.

[0006] A computing device configured to perform object detection in the open world includes a processing device configured to execute instructions stored in a memory to receive object data, the object data including embedding data corresponding to embeddings for known objects in a first input image, projecting the embeddings into a hyperbolic embedding space containing embeddings in a plurality of categories of objects, each containing one or more classes of objects, regularizing the projected embeddings in the hyperbolic embedding space by moving each of the projected embeddings closer to embeddings in the same category of the plurality of categories and farther away from embeddings in different categories of the plurality of categories, receiving an unmatched query corresponding to an object in a second input image,and generating, based on the hyperbolic embedding space containing the regularized embeddings, an output signal indicating whether the object in the second input image corresponds to an unknown object in one of the classes of objects.,

[0007] According to further features, the object data includes a set of bounding boxes and class labels for the known objects in the first input image. The first input image corresponds to a training image. The processing device is further configured to execute instructions stored in a memory to generate a hyperbolic contrastive loss corresponding to the hyperbolic embedding space, wherein regularizing the projected embeddings comprises regularizing the projected embeddings based on the hyperbolic contrastive loss. To regularize the projected embeddings, the processing device is further configured to execute instructions stored in a memory to determine respective hyperbolic averages of embeddings in each of the classes of objects.To generate the output signal, the processing device is further configured to execute instructions stored in a memory to determine a threshold distance based on the hyperbolic averages and to generate the output signal based in part on the threshold distance. To generate the output signal, the processing device is further configured to execute instructions stored in a memory to determine whether the unmatched sample is less than the threshold distance of at least one of the hyperbolic averages.

[0008] A computer-controlled machine includes at least one sensor configured to generate a first input image and a second input image, a control system configured to receive object data including embedding data corresponding to embeddings for known objects in a first input image, project the embeddings into a hyperbolic embedding space containing embeddings in a plurality of categories of objects, each containing one or more classes of objects, regularize the projected embeddings in the hyperbolic embedding space by moving each of the projected embeddings closer to embeddings in the same category of the plurality of categories and farther away from embeddings in different categories of the plurality of categories, receive an unmatched query corresponding to an object in a second input image, and generate, based on the hyperbolic embedding space,containing the regularized embeddings, an output signal indicating whether the object in the second input image corresponds to an unknown object in one of the classes of objects, and an actuator configured to control an operation of the computer-controlled machine in response to the output signal.

[0009] According to further features, the computer-controlled machine further includes a memory that stores data corresponding to the hyperbolic embedding space. The control system is further configured to generate a hyperbolic contrastive loss corresponding to the hyperbolic embedding space, wherein regularizing the projected embeddings comprises regularizing the projected embeddings based on the hyperbolic contrastive loss. To regularize the embeddings, the control system is further configured to determine respective hyperbolic averages of embeddings in each of the classes of objects. To generate the output signal, the control system is further configured to determine a threshold distance based on the hyperbolic averages and to determine whether the unmatched sample is less than the threshold distance of at least one of the hyperbolic averages.The computer-controlled machine contains an autonomous robot. Brief description of the drawings Fig. 1 generally illustrates a system for training a neural network according to the principles of the present disclosure. Fig. 2 generally illustrates a computer-implemented method for training and using a neural network according to the principles of the present disclosure. Fig. 3A generally illustrates an audio data labeling system according to the principles of the present disclosure. Fig. 3B generally illustrates a portion of a data acquisition system according to the principles of the present disclosure. Fig. 3C generally illustrates an alternative audio data labeling system according to the principles of the present disclosure. Fig. 4A is a functional block diagram of an exemplary open world hyperbolic system according to the principles of the present disclosure. Fig. 4B is a table illustrating example performance of a dataset according to principles of the present disclosure with respect to other example datasets for various tasks. Fig. 4C illustrates T-distributed stochastic neighbor (t-SNE) graphical representations illustrating a learned embedding achieved using hyperbolic distance according to principles of the present disclosure. Fig. 4D is a table illustrating an example ablation analysis that quantifies the impact of individual components of techniques of the present disclosure. Fig. 4E illustrates example class distance heat maps quantifying hyperbolic distance according to the principles of the present disclosure. Fig. 4F illustrates an example distance heat map quantifying a hyperbolic distance between known elements and unknown elements according to the principles of the present disclosure. Fig. 4G depicts graphical t-SNE representations of learned embeddings using cosine and hyperbolic distance metrics according to the principles of the present disclosure. Fig. 4H illustrates steps of an exemplary open world hyperbolic method according to the principles of the present disclosure. Fig. 5 illustrates a schematic diagram of an interaction between a computer-controlled machine and a control system according to the principles of the present disclosure. Fig. Figure 6 shows a schematic diagram of the control system of Fig. 5 configured to control a vehicle, which may be a semi-autonomous vehicle, a fully autonomous vehicle, a semi-autonomous robot, or a fully autonomous robot, in accordance with the principles of the present disclosure. Fig. Figure 7 shows a schematic diagram of the control system of Fig. 5 configured to control a manufacturing machine such as a punch, a cutter, or a gun drill of a manufacturing system such as part of a production line. Fig. Figure 8 shows a schematic diagram of the control system of Fig. 5, which is configured to control a power tool such as a drill or a screwdriver, which has an at least partially autonomous mode. Fig. Figure 9 shows a schematic diagram of the control system of Fig. 5 configured to control an automated personal assistant. Fig. Figure 10 shows a schematic diagram of the control system of Fig. 5 configured to control a control system such as an access control system or a surveillance system. Fig. Figure 11 shows a schematic diagram of the control system of Fig. 5 configured to control an imaging system, e.g., an MRI device, an X-ray imaging device, or an ultrasound device. Detailed description

[0010] Embodiments of the present disclosure are described herein. However, it should be understood that the disclosed embodiments are merely examples, and further embodiments may take various and alternative forms. The figures are not necessarily to scale; some features may be exaggerated or minimized to show details of particular components. Therefore, specific structural and functional details disclosed herein are not to be interpreted as limiting, but merely as representative bases for teaching one skilled in the art to variously employ the embodiments.As will be understood by those skilled in the art, various features illustrated and described with reference to any of the figures may be combined with features illustrated in one or more other figures to produce embodiments not explicitly illustrated or described. The combinations of features illustrated provide representative embodiments for typical application. However, various combinations and modifications of the features consistent with the teachings of this disclosure may be desirable for particular applications or implementations.

[0011] As used herein, "a," "an," and "the" refer to both singular and plural references unless the context clearly indicates otherwise. For example, "a processor" programmed to perform various functions refers to one processor programmed to perform each function individually, or more than one processor programmed together to perform each of the various functions.

[0012] Open-world object detection (OWOD) involves detecting both known (previously seen and classified during training) and unknown (previously unseen and classified and / or outside of known classes) objects. However, the level of "unfamiliarity" varies considerably depending on the context. For example, a tree is typically considered part of the background in a self-driving scene, but can also be essential in a household context.

[0013] The OWOD framework incorporates both detection tasks and continuous learning. However, a key challenge in this framework is the lack of annotations for unknown objects, which leads to biases toward known labels and potential confusion between unknown elements and the background. This bias significantly complicates the accurate identification of unknown objects and represents a major hurdle in the detection process.

[0014] Previous attempts to solve this problem include various techniques such as relying on a learned objectness score to relabel the background as possible unknowns, grouping classes to better isolate unknowns, and a decoupled approach where the classification and localization heads are separated.

[0015] The task is to remove label information and instead include the shared features that make labeled elements relevant as objects.

[0016] However, these techniques fail to address a key problem: defining what constitutes an "unknown" object. Currently, no clear definition or prior knowledge is available to effectively distinguish unknown objects from the background, and interpretation varies considerably depending on the context. For example, in a driving scene, "debris on the road" might be considered an unknown object relevant to a task. Conversely, in a camera surveillance context, the "debris on the road" might be perceived as part of the background. Without considering the context, prior techniques can only learn to differentiate knowns from unknowns based on low-level features such as texture or shape.As a result, modeling of hierarchical structures and similarities between known and unknown elements at both the image level and the data set level is not performed.

[0017] Accordingly, systems and methods according to the principles of the present disclosure define a hierarchical structural relationship between the objects to be detected and the known elements or objects. This hierarchy is characterized by classes that share the same semantic context and belong to the same category (e.g., vehicles, animals, electronics, etc.). Such hierarchical relationships enable the retrieval of common features and facilitate the discovery of unknown objects. For example, a model trained on objects related to driving scenes may adequately detect stop signs or traffic lights, but is not expected to be able to detect unrelated objects such as a bed or other furniture.

[0018] Therefore, as implemented in systems and methods of the present disclosure, external or context-dependent information is already embedded in the known classes. In other words, there is a semantic or latent structural relationship between the known and unknown elements to be discovered. The systems and methods described herein learn and model a hierarchical representation of known elements using a regularizer configured according to the principles of the present disclosure (e.g., a SuperClass regularizer). Leveraging the learned hierarchical representation enables detection of unknown objects using adaptive relabeling (a similarity-distance-based relabeling module, circuitry, etc.).

[0019] In this way, hierarchical relationships between elements are learned to effectively use the representation of known objects for the detection of unknown elements. For example, elements belonging to the same family (or category) may be closer to each other while being farther away from different families (e.g., closer to other animals while being farther away from vehicles). To accommodate this structural hierarchy, the systems and methods described here compute and operate according to a hyperbolic distance (which maps to hierarchical latent structures such as graphs or trees and acts as a natural distance metric). This mapping in hyperbolic space has the desirable property of capturing the affinity between unknown elements and known elements, thereby improving the detection of unknown objects.

[0020] Fig. 1 shows an example system 100 for training a neural network (e.g., an ML model). The system 100 may be configured (and / or include circuitry configured) to implement the systems and methods of the present disclosure, which is described in more detail below. The system 100 may include an input interface for accessing training data 102 for the neural network. For example, as shown in Fig. 1, the input interface may be formed by a data storage interface 104 that can access the training data 102 from data storage 106. For example, the data storage interface 104 may be a memory interface or a persistent storage interface, e.g., a hard disk or SSD interface, but may also be a personal, local, or wide area network interface, such as a Bluetooth, Zigbee, or Wi-Fi interface, or an Ethernet or fiber optic interface. The data storage 106 may be an internal data storage of the system 100, such as a hard disk or SSD, but may also be an external data storage, e.g., a network-accessible data storage.

[0021] In certain embodiments, the data store 106 may further include a data representation 108 of an untrained version of the neural network, accessible by the system 100, from the data store 106. However, it is recognized that the training data 102 and the data representation 108 of the untrained neural network may each be accessed from a different data store, e.g., via a different subsystem of the data store interface 104. Each subsystem may be of a type as described above for the data store interface 104.

[0022] In certain embodiments, the data representation 108 of the untrained neural network may be generated internally by the system 100 based on design parameters for the neural network and therefore does not need to be explicitly stored in the data store 106. The system 100 may further include a processor subsystem 110 that may be configured, during operation of the system 100, to provide an iterative function as a replacement for a stack of layers of the neural network to be trained. Here, respective layers of the stack of layers being replaced may share weights with each other and may receive as input an output of a previous layer or, for a first layer of the stack of layers, an initial activation and a portion of the input of the stack of layers.

[0023] The processor subsystem 110 may be further configured to iteratively train the neural network using the training data 102. Here, an iteration of the training by the processor subsystem 110 may include a forward propagation portion and a backward propagation portion. The processor subsystem 110 may be configured to perform the forward propagation portion, among further operations defining the forward propagation portion that may be performed, by determining an equilibrium point of the iterative function at which the iterative function converges to a fixed point, wherein determining the equilibrium point comprises using a numerical root-search algorithm to find a root solution for the iterative function minus its input, and by providing the equilibrium point as a substitute for an output of the stack of layers in the neural network.

[0024] The system 100 may further comprise an output interface for outputting a data representation 112 of the trained neural network. This data may also be referred to as trained model data 112. For example, as also shown in Fig. 1, the output interface may be formed by the data storage interface 104, wherein the interface in these embodiments is an input / output interface ('IO' interface) by means of which the trained model data 112 may be stored in the data storage 106. For example, the data representation 108 defining the 'untrained' neural network may be at least partially replaced by the data representation 112 of the trained neural network during or after training, in that the parameters of the neural network, such as weights, hyperparameters, and other types of neural network parameters, may be designed to reflect the training on the training data 102. This is illustrated in Fig. 1 also by reference numerals 108, 112 with reference to the same data set in data storage 106. In certain embodiments, data representation 112 may be stored separately from data representation 108, which defines the 'untrained' neural network. In certain embodiments, the output interface may be separate from data storage interface 104, but may generally be of a type as described above for data storage interface 104.

[0025] Fig. 2 illustrates a data annotation / enhancement system 200 configured (and / or including circuitry configured) to implement a system for annotating and / or annexing data. The data annotation system 200 may include at least one computing system 202 configured to implement all or portions of the systems and methods of the present disclosure, which are explained in more detail below. The computing system 202 may include at least one processor 204 operatively connected to a memory unit 208. The processor 204 may include one or more integrated circuits implementing the functionality of a central processing unit (CPU) 206. The CPU 206 may be a commercially available processing unit implementing an instruction set, such as one of the x86, ARM, Power, or MIPS instruction set families.Various components of system 200 may be implemented with the same or different circuitry.

[0026] During operation, CPU 206 may execute stored program instructions retrieved from memory unit 208. The stored program instructions may include software that controls the operation of CPU 206 to perform the operation described herein. In certain embodiments, processor 204 may be a system-on-chip (SoC) that integrates functionality of CPU 206, memory unit 208, a network interface, and input / output interfaces into a single integrated device. Computing system 202 may implement an operating system for managing various aspects of operation.

[0027] The storage unit 208 may include volatile memory and non-volatile memory for storing instructions and data. The non-volatile memory may include solid-state memory such as NAND flash memory, magnetic and optical storage media, or any other suitable data storage device that retains data when the computing system 202 is deactivated or loses power. The volatile memory may include static and dynamic random access memory (RAM) that stores program instructions and data. For example, the storage unit 208 may include a machine learning model 210 (e.g., used in Fig. 2 as the ML model 210) or an algorithm, a training dataset 212 for the machine learning model 210, a source raw dataset 216, etc.

[0028] Computing system 202 may include a network interface device 222 configured to provide communication with external systems and devices. For example, network interface device 222 may include a wired and / or wireless Ethernet interface as defined by the 802.11 family of standards of the Institute of Electrical and Electronics Engineers (IEEE). Network interface device 222 may include a mobile communications interface for communicating with a cellular network (e.g., 3G, 4G, 5G). Network interface device 222 may be further configured to provide a communications interface to an external network 224 or a cloud.

[0029] The external network 224 may be referred to as the "World Wide Web" or the Internet. The external network 224 may establish a standard communication protocol between computing devices. The external network 224 may enable information and data to be easily exchanged between computing devices and networks. One or more servers 230 may be in communication with the external network 224.

[0030] Computing system 202 may include an input / output (I / O) interface 220 that may be configured to provide digital and / or analog inputs and outputs. Input / output interface 220 may include additional serial interfaces for communicating with external devices (e.g., a Universal Serial Bus (USB) interface).

[0031] Computing system 202 may include a human-machine interface (HMI) device 218, which may include any device that enables system 200 to receive control input. Examples of input devices may include human interface inputs such as keyboards, mice, touch screens, voice input devices, and other similar devices. Computing system 202 may include a display device 232. Computing system 202 may include hardware and software for outputting graphic and text information to display device 232. Display device 232 may include an electronic display screen, a projector, a printer, or any other suitable device for displaying information to a user or operator.The computing system 202 may be further configured to enable interaction with remote HMI and remote display devices via the network interface device 222.

[0032] System 200 may be implemented using one or more computing systems. While the example illustrates a single computing system 202 implementing all of the described features, it is intended that various features and functions may be implemented separately and by multiple computing units communicating with each other. The particular system architecture chosen may depend on a variety of factors.

[0033] The system 200 may implement a machine learning model 210 configured to analyze the source raw data set 216. For example, the CPU 206 and / or other circuitry may implement the machine learning model 210. The source raw data set 216 may include raw or unprocessed sensor data, which may represent an input data set for a machine learning system. The source raw data set 216 may include a video, video segments, images, audio, text-based information, and raw or partially processed sensor data (e.g., a radar map of objects). In certain embodiments, the machine learning model 210 may be a deep learning or neural network algorithm configured to perform a predetermined function.For example, a neural network algorithm may be configured to identify events or objects in video segments based on audio data.

[0034] Computer system 200 may store the training dataset 212 for the machine learning model 210. The training dataset 212 may represent a set of previously constructed data for training the machine learning model 210. For example, according to the present disclosure, the training dataset 212 may include a plurality of automatically collected ground truth measurements and associated data. The training dataset 212 may be used by the machine learning model 210 to learn weighting factors associated with a neural network algorithm. The training dataset 212 may include a set of source data having corresponding outcomes or results that the machine learning model 210 attempts to duplicate through the learning process.

[0035] The machine learning model 210 may operate in a learning mode using the training dataset 212 as input. The machine learning model 210 may be executed for a number of iterations using the data from the training dataset 212. With each iteration, the machine learning model 210 may update internal weighting factors based on the results obtained. For example, the machine learning model 210 may compare output results (e.g., annotations) with those contained in the training dataset 212. Because the training dataset 212 contains the expected results, the machine learning model 210 may determine whether the performance is acceptable. After the machine learning model 210 has achieved a predetermined performance level (e.g.,Once the machine learning model 210 has achieved a high level of agreement (e.g., 100% agreement with the results associated with the training dataset 212), the machine learning model 210 may be executed using data not in the training dataset 212. The trained machine learning model 210 may be applied to new datasets to generate annotated data.

[0036] The machine learning model 210 may be configured to identify a particular feature in the source raw data 216. The source raw data 216 may contain a plurality of instances or an input data set for which annotation results are desired (e.g., a video stream or video segment containing audio data). By way of example only, the machine learning model 210 may be configured to identify objects or events in a video segment based on audio data and annotate the events. The machine learning model 210 may be programmed to process the source raw data 216 to identify the presence of the particular features. The machine learning model 210 may be configured to identify a feature in the source raw data 216 as a predetermined feature. The source raw data 216 may be derived from a variety of sources.For example, the source raw data 216 may be actual input data collected by a machine learning system. The source raw data 216 may be machine-generated for testing the system. As an example, the source raw data 216 may include raw video and / or audio data from a camera, audio data from a microphone, etc.

[0037] In one example, machine learning model 210 may process raw source data 216 and output video and / or audio data containing one or more indications of an identified event. Machine learning model 210 may generate a confidence level or confidence factor for each generated output. For example, a confidence value that exceeds a predetermined high-confidence threshold may indicate that machine learning model 210 is confident that the identified event (or feature) corresponds to the particular event. A confidence value that is less than a low-confidence threshold may indicate that machine learning model 210 has some uncertainty that the particular feature is present.

[0038] As in Fig. 3A and Fig. 3B, a system 300 may include an image capture device (e.g., a video capture device) 302, an audio capture device 304, and the computing system 202. The system may receive video stream data associated with a data capture environment from the image capture device 302. The system 202 may be configured to perform video object detection to identify one or more objects in corresponding frames of the video stream data. The system 202 may receive audio stream data corresponding to at least a portion of the video stream data from the audio capture device 304. The audio capture device 304 may include one or more microphones 306 or other suitable audio capture devices. The systems and methods described herein may be configured to perform video object detection using an output of at least a first machine learning model (such asthe machine learning model 210 or another suitable machine learning model configured to provide an output including one or more object or event detection predictions) to label at least some objects of the video stream data and / or the audio stream data.

[0039] The system 202 may calculate (e.g., using at least one probabilistic function or another suitable technique or function) at least one offset value for at least a portion of the audio stream data corresponding to at least one tagged object of the video stream data based on at least one data acquisition characteristic. The system 202 may synchronize at least a portion of the video stream data with the portion of the audio stream data corresponding to the at least one tagged object of the video stream data using at least the at least one offset value.The at least one data capture characteristic may include one or more properties of the at least one image capture device, one or more properties of the at least one audio capture arrangement, one or more properties corresponding to a location of the at least one image capture device relative to the at least one audio capture arrangement, one or more properties corresponding to a movement of an object in the video stream data, one or more other suitable data capture characteristics, or a combination thereof.

[0040] The system 202 may label at least the portion of the audio stream data corresponding to the at least one labeled object of the video stream data using one or more labels of the labeled objects of the video stream data and the at least one offset value. Each respective label may include an event type, an event start indicator, and an event end indicator. The system 202 may generate training data using at least a portion of the labeled portion of the audio stream data. The system 202 may train a second machine learning model using the training data. The system 202 may detect, using the second machine learning model, one or more sounds associated with audio data provided as input to the second machine learning model.The second machine learning model may include any suitable machine learning model and may be configured to implement any suitable function such as those described herein with respect to . Fig. 4- Fig. 11.

[0041] In certain embodiments, as in Fig. 3C, the computing system 202 may be configured to label audio data based on sensor data received from one or more sensors, such as those described herein, or any other suitable sensor or combination of sensors. The system 202 may receive audio stream data associated with a data capture environment from the audio capture assembly 354 or any suitable audio capture device, such as one or more of the microphones 306 or any other suitable audio capture device. It should be understood that the audio capture assembly 354 may include features similar to those of the audio capture assembly 304 and may include any suitable number of audio capture devices. The system 202 may receive audio stream data from at least one sensor (e.g., such asThe audio pickup assembly 354 may receive sensor data associated with the data pickup environment from a sensor 352 that is asynchronous with respect to the audio pickup assembly 354. The sensor 354 may include an induction coil, a radar sensor, a LiDAR sensor, a sonar sensor, an image pickup device, and / or another suitable sensor, or a combination thereof. The audio pickup assembly 354 may be located remotely from the sensor 354, proximate to the sensor 354, or in any suitable relationship to the sensor 354.

[0042] The system 202 may identify at least some events in the sensor data using an output of at least a first machine learning model, such as the machine learning model 210 or another suitable machine learning model. The machine learning model 210 may be configured to provide an output including one or more event detection predictions based on the sensor data. The system 202 may synchronize at least a portion of the sensor data associated with the portion of the audio stream data corresponding to the at least one event of the sensor data. The system 202 may label at least the portion of the audio stream data corresponding to the at least one event of the sensor data using one or more labels extracted for respective events of the sensor data value.Each respective label may include an event type, an event start indicator, and an event end indicator. The system 202 may generate training data using at least a portion of the labeled portion of the audio stream data. The system 202 may train a second machine learning model using the training data. The system 202 may detect, using the second machine learning model, one or more sounds associated with audio data provided as input to the second machine learning model. The second machine learning model may include any suitable machine learning model and may be configured to implement any suitable function, such as those described herein with respect to. Fig. 4- Fig. 11.

[0043] Any of the systems described in more detail above and / or below may implement systems and methods of the present disclosure to perform an OWOD. Systems and methods of the present disclosure may be referred to as hyperbolic open-world systems and methods ("Hyp-OW" systems and methods).

[0044] Fig. 4A is a functional block diagram of an exemplary Hyp-OW system 400 according to the present disclosure. The system 400 includes one or more convolutional neural networks, such as a deformable transformer 402, configured to receive and process inputs corresponding to an observed scene (e.g., an input image 404). For example, the deformable transformer 402 implements an object detection model that performs a deformable convolution on specific regions of the input image 404. An output 406 of the deformable transformer is a set of bounding boxes and class labels for objects detected in the input image 404, which are provided to a classification system 408 according to the present disclosure.

[0045] The classification system 408 includes a classifier 410 (e.g., a classifier configured to classify input signals into one or more labels using a machine learning (ML) algorithm such as a neural network), a hyperbolic contrastive loss module 412 configured to learn a hierarchical structure representation for each class, and a SuperClass regularization module 414 configured to regularize classes such that classes belonging to the same category are closer to each other in an embedding space (e.g., a hyperbolic embedding space) while being farther away from classes of different categories. An adaptive relabeling module 416 leverages the learned hierarchical structure to relabel proposed candidates as unknowns if they meet a certain criterion, as described in more detail below.

[0046] For example, different embeddings 418 (e.g., vector representations that may be referred to as embedding data) of images or image regions are positioned in an embedding space 420. To ensure that classes belonging to the same category (e.g., vehicles, animals, electronics, etc.) are closer to each other in the embedding space 420 and farther away from classes of different categories, new (unmatched) embeddings may be moved closer to additional classes in the same category. The adaptive relabeling module 416 receives known elements and unknown candidates (e.g., from an input image 422) and may selectively relabel embeddings (e.g., based on a comparison between a distance d of the unmatched embedding from a predetermined or calculated position (e.g., a centroid position) for the category in the embedding space 420 and a threshold distance δ).For example, if the distance d is less than or equal to the threshold distance δ, the embedding can be relabeled. Conversely, if the distance d is greater than the threshold distance δ, the embedding does not need to be relabeled.

[0047] In this way, the Hyp-OW system 400 is configured to effectively capture hierarchical relationships between known and unknown classes, resulting in improved object detection performance.

[0048] In one embodiment, each embedding is projected into a hyperbolic space (i.e., a hyperbolic embedding space) with an exponential mapping according to the following: exp0c(x)=tanh(c‖x‖2)xc‖x‖.

[0049] For each embedding query q, a corresponding projection z into hyperbolic space is defined as: z=exp0c(q).

[0050] The hyperbolic contrastive loss module 412 is configured to learn the hierarchical structure representation for each class to bring embeddings of the same classes closer together while pushing each class away from embeddings of different classes. For example, z corresponds to i c a query that is matched to a class c ∈ K. For each query z i c a positive example is sampled from a buffer M. The query z i c , where i ∈ A = {B ∪ M}, is the query embedding of class c, which can come either from a set B or the buffer M. A sample z i+ corresponds to a positive sample value, while z i-corresponds to a negative sample. To attract embeddings of the query and the positive sample while repelling embeddings of negative samples, a hyperbolic contrastive loss can therefore be written as: Lhyp=∑{i∈A,c∈K}logexp−dhyp(zic,zi+)τ1∑i−∈A / {i+,i}−dhyp(zic,zi−)τ1, where τ1 is the temperature, dhyp(x,zk)=2carctanh(c‖−x⊕cy‖) and x⊕cy=(1+2c<x,y> +c‖y‖2)x+(1−c‖x‖2)y(1+2c<x,y> c2‖x‖2‖y‖2).

[0051] The SuperClass regularization module 414 is configured to bring classes from the same category closer together and push them away from classes belonging to different categories. For a set of class indices S p , which belong to a category p = 1...P, in order to approximate an embedding of category p, a hyperbolic average of each embedding {ziC}, i ∈ M of class c ∈ Sp according to Hp=∑i∈M,c∈Spγizic∑i∈Mγi calculated, where γi=11−c‖x‖2 corresponds to the Lorentz factors.

[0052] For each embedding query z i c a category embedding H p (c ∈ S p ) is sampled from the buffer M and the same hyperbolic contrastive loss is used according to the following: Lhyp=∑{i∈A,c∈Sp}logexp−dhyp(zic,Hp)τ2∑k≠p,i∉Sk−dhyp(zic,Hk)τ2

[0053] The adaptive relabeling module 416 is configured to implement an adaptive relabeling scheme for queries such as m from the incoming set B, which were matched with a ground truth label, and queries z u that have not been matched with a ground-truth label (referred to as “proposal candidates”). Z ccorresponds to a hyperbolic average of each class c from the buffer M and represents a centroid in the hyperbolic embedding space for each class. Distances between matched queries z m and the centers of gravity are calculated to achieve a maximum distance δ B to be identified according to the following: δB=maxm∈B,c∈Kdhyp(zm,Zc).

[0054] Distances between each unmatched query z u and all focus areas Z c are calculated and a minimum distance for each z u is selected. If the minimum distance is less than δ B , it can be assumed that the unmatched query probably contains an unknown element. Accordingly, any unmatched query can u be relabeled in response to the following condition being true: minc∈K dhyp(zu,Zc)≤δB.

[0055] The unmatched queries can then be relabeled as unknowns and forwarded for classification.

[0056] A final loss L can be calculated according to the following: L = L cls + L bbox + αL hyp + βL reg , where the coefficients α and β weight the hyperbolic contrastive loss and the SuperClass regularizer, respectively.

[0057] Fig. Figure 4B shows a table 424 illustrating example performance of a dataset ("Hyp-OW") of the systems and methods of the present disclosure relative to other example datasets for various tasks. As shown, the datasets can be evaluated for low, medium, and high difficulty levels. Evaluation metrics include U-Recall (e.g., unknown recall shown as a percentage of unknown objects detected) for unknown detection and mAP for known detection. Hyp-OW shows consistent improvement across other techniques for both unknown and known detection at all difficulty levels.

[0058] Fig. 4C shows graphical representations of T-distributed stochastic neighbors (t-SNE graphical representations) 426 and 428 illustrating a learned embedding achieved using hyperbolic distance with (at 426) and without (at 428) use of the SuperClass regularization module 414 according to the present disclosure. Each class is represented based on a corresponding category. As shown, using the SuperClass regularization module 414 as described herein enforces different grouping of classes belonging to the same categories, enhancing a hierarchical structure and effectively separating the embeddings of different categories.

[0059] Fig. 4D shows a table 432 illustrating an example ablation analysis quantifying the impact of individual components of Hyp-OW techniques of the present disclosure (e.g., results obtained using a cosine distance, results obtained without using the SuperClass regularization module 414, and results obtained without using adaptive relabeling). The results show that the adaptive relabeling module 416 improves the accuracy of known detection (mAP), while the SuperClass regularization module 414 improves the accuracy of unknown detection (U-recall). Conversely, using a cosine distance (i.e., instead of hyperbolic distance) results in decreased unknown detection accuracy.

[0060] Fig. 4E shows example class distance heat maps 436 and 438 quantifying a hyperbolic distance (e.g., ranging from 1.0 to 1.5) between category embeddings (e.g., vehicle, animal, outdoor, furniture) and individual classes (e.g., bicycle, passenger car, motorcycle, airplane, etc.), where a lighter shading indicates a smaller distance (and greater similarity) and a darker shading indicates a larger distance (and less similarity). Results obtained using the SuperClass regularization module 414 are shown at 436, and results obtained without using the SuperClass regularization module 414 are shown at 438. The heat map 436 illustrates that using the SuperClass regularization module 414 improves both within-category and between-category distances.For example, as shown, classes in groupings 440, 442, 444, and 446 are closer to respective categories and farther from other categories. Accordingly, the use of the SuperClass regularization module 414 separates different categories from each other in hyperbolic space and groups classes in the same category more closely together in hyperbolic space. In contrast, without the use of the SuperClass regularization module 414, as shown at 438, distances between categories in hyperbolic space appear less pronounced, as indicated by a weaker shading intensity.

[0061] Fig. Figure 4F shows an example distance heat map 450 quantifying a hyperbolic distance between known items and unknown items for a hierarchical partitioning task (e.g., Task 1), where a lighter shade indicates a smaller distance (and greater similarity) and a darker shade indicates a larger distance (and less similarity). As shown, a smaller distance is observed between known and unknown items belonging to the same category (as shown by boxes 452 in some examples). Accordingly, unknown items can be effectively detected and labeled based on their similarity to known items (e.g., by the adaptive relabeling module 416).

[0062] Fig. Figure 4G shows graphical t-SNE representations of 456 and 458 learned embeddings using two different distance metrics: cosine distance (at 456) and hyperbolic distance (at 458). Using hyperbolic distance deforms the embeddings in a hierarchically structured manner such that classes belonging to the same category exhibit more cohesive grouping (i.e., improved clustering) and embeddings of different categories are effectively separated.

[0063] Fig. 4H illustrates steps of an exemplary hyperbolic open-world object detection method 460 according to the principles of the present disclosure. At 462, method 460 receives object data corresponding to an input image. For example, method 460 receives a set of bounding boxes and class labels for objects detected in an input image that correspond to known element embeddings. At 464, method 460 projects the embeddings into hyperbolic space. At 466, method 460 determines a hyperbolic contrastive loss, as described above.

[0064] At 468, the method 460 regularizes the embeddings in hyperbolic space. As used herein, "regularizing" refers to regularizing classes such that classes belonging to the same category are closer to each other in an embedding space (e.g., a hyperbolic embedding space), while they are farther away from classes of different categories (e.g., using the SuperClass regularization module 414, as described above). For example, regularizing the embeddings according to the principles of the present disclosure includes determining a hyperbolic average of embeddings in a given class.

[0065] At 470, method 460 performs adaptive relabeling (e.g., using adaptive relabeling module 416) on proposed unknown candidates (e.g., unmatched queries) from an input image. For example, proposed unknown candidates that are less than a predetermined distance (e.g., the threshold distance described above) from at least one computed centroid of all categories are relabeled as unknowns.

[0066] As described above, systems and methods according to the principles of the present disclosure highlight a semantic relationship between known and unknown elements. For example, a Hyp-OW dataset according to the principles of the present disclosure is configured to define a high level of hierarchy among elements that highlights the semantic relationship. As described above, the Hyp-OW systems and methods perform the following: (i) applying a hyperbolic contrastive loss to learn and model the hierarchical structure by moving embeddings of the same classes closer together and moving embeddings away from embeddings of different classes (e.g., using the hyperbolic contrastive loss module 412); (ii) moving classes from the same category closer together while pushing classes away from classes of different categories (e.g.,using the SuperClass regularization module 414); and (iii) implementing adaptive relabeling to use a semantic similarity distance to dynamically relabel proposed candidates as unknowns to leverage the learned hierarchical structure (e.g., using the adaptive relabeling module 416).

[0067] Fig. 5- Fig. 11 illustrate example systems and devices that may implement OWOD systems and methods according to the present disclosure. Fig. 5 illustrates a schematic diagram of an interaction between a computer-controlled machine 500 and the control system 502. The computer-controlled machine 500 includes an actuator 504 and a sensor 506. The actuator 504 may include one or more actuators, and the sensor 506 may include one or more sensors. The sensor 506 is configured to sense a state of a computer-controlled machine 500. The sensor 506 may be configured to encode the sensed state into sensor signals 508 and send sensor signals 508 to control a system 502. Non-limiting examples of a sensor 506 include video, radar, LiDAR, ultrasonic, and motion sensors. In certain embodiments, the sensor 506 is an optical sensor configured to capture optical images of an environment proximate a computer-controlled machine 500.

[0068] The control system 502 is configured to receive sensor signals 508 from the computer-controlled machine 500. As set forth below, the control system 502 may be further configured to calculate actuator control instructions 510 that depend on the sensor signals and send actuator control instructions 510 to the actuator 504 of the computer-controlled machine 500.

[0069] As in Fig. 5, a control system 502 includes the receiving unit 512. The receiving unit 512 may be configured to receive sensor signals 508 from the sensor 506 and convert the sensor signals 508 into input signals x. In an alternative embodiment, sensor signals 508 are received directly as input signals x without the receiving unit 512. Each input signal x may be a portion of each sensor signal 508. The receiving unit 512 may be configured to process each sensor signal 508 to generate each input signal x. The input signal x may include data corresponding to an image recorded by the sensor 506.

[0070] The control system 502 includes a classifier 514. The classifier 514 may be configured to classify input signals x into one or more labels using a machine learning (ML) algorithm, such as a neural network. The classifier 514 is configured to be parameterized by parameters such as those described above (e.g., the parameter θ). The parameter θ may be stored in and provided by a non-volatile memory 516. The classifier 514 is configured to determine output signals y from input signals x. Each output signal y includes information assigning one or more labels to each input signal x. The classifier 514 may send output signals y to the conversion unit 518. The conversion unit 518 is configured to convert output signals y into actuator control instructions 510.The control system 502 is configured to send actuator control instructions 510 to the actuator 504, which is configured to actuate a computer-controlled machine 500 in response to the actuator control instructions 510. In certain embodiments, the actuator 504 is configured to actuate a computer-controlled machine 500 directly based on output signals y.

[0071] Upon receipt of actuator control instructions 510 by actuator 504, actuator 504 is configured to perform an action corresponding to the related actuator control instruction 510. Actuator 504 may include control logic configured to convert actuator control instructions 510 into a second actuator control instruction used to control actuator 504. In one or more embodiments, actuator control instructions 510 may be used to control a display device instead of, or in addition to, an actuator.

[0072] In certain embodiments, a control system 502 includes the sensor 506 instead of, or in addition to, the computer-controlled machine 500 that includes the sensor 506. The control system 502 may also include the actuator 504 instead of, or in addition to, the computer-controlled machine 500 that includes the actuator 504.

[0073] As in Fig. 5, a control system 502 further includes a processor 520 and a memory 522. The processor 520 may include one or more processors. The memory 522 may include one or more storage devices. The classifier 514 (e.g., ML algorithms) of one or more embodiments may be implemented by the control system 502, which includes the non-volatile memory 516, the processor 520, and the memory 522.

[0074] Non-volatile memory 516 may include one or more persistent data storage devices, such as a hard drive, an optical drive, a tape drive, a non-volatile solid-state device, cloud storage, or any other device capable of persistently storing information. Processor 520 may include one or more devices selected from high-performance computing (HPC) systems containing high-performance cores, microprocessors, microcontrollers, digital signal processors, microcomputers, central processing units, field-programmable gate arrays, programmable logic devices, state machines, logic circuits, analog circuits, digital circuits, or other devices that manipulate signals (analog or digital) based on computer-executable instructions located in memory 522.The memory 522 may include a single memory device or multiple memory devices including, but not limited to, random access memory (RAM), volatile memory, non-volatile memory, static random access memory (SRAM), dynamic random access memory (DRAM), flash memory, cache, or any other device capable of storing information.

[0075] Processor 520 may be configured to read into memory 522 and execute computer-executable instructions located in non-volatile memory 516 and embodying one or more anomaly detection methodologies of one or more embodiments. Non-volatile memory 516 may include one or more operating systems and applications. Non-volatile memory 516 may store compiled and / or interpreted computer programs created using a variety of programming languages ​​and / or technologies, including, without limitation, Java, C, C++, C#, Objective C, Fortran, Pascal, Java Script, Python, Perl, and PL / SQL, either alone or in combination.

[0076] Upon execution by processor 520, the computer-executable instructions of non-volatile memory 516 may cause a control system 502 to implement one or more of the anomaly detection methodologies disclosed herein. Non-volatile memory 516 may also contain data supporting the functions, features, and processes of the one or more embodiments described herein.

[0077] The program code embodying the algorithms and / or methodologies described herein may be distributed individually or collectively as a program product in a variety of different forms. The program code may be distributed using a computer-readable storage medium containing computer-readable program instructions for causing a processor to perform aspects of one or more embodiments. Computer-readable storage media, which are inherently non-transitory, may include volatile and non-volatile, and removable and non-removable tangible media implemented in any method or technology for storing information, such as computer-readable instructions, data structures, program modules, or other data.Computer-readable storage media may further include RAM, ROM, erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other solid-state storage technologies, a portable compact read-only memory (CD-ROM) or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store the desired information and that can be read by a computer. Computer-readable program instructions may be downloaded to a computer, another type of programmable data processing apparatus, or another device from a computer-readable storage medium, or to an external computer or external storage device via a network.

[0078] Computer-readable program instructions stored on a computer-readable medium may be used to instruct a computer, other types of programmable data processing apparatus, or other devices to operate in a particular manner, such that the instructions stored on the computer-readable medium produce an article of manufacture including instructions that implement the functions, acts, and / or operations specified in the flowcharts or diagrams. In certain alternative embodiments, the functions, acts, and / or operations specified in the flowcharts and diagrams may be reordered, serialized, and / or temporally overlapped consistent with one or more embodiments.Additionally, consistent with one or more embodiments, any flowcharts and / or diagrams may include more or fewer nodes or blocks than those illustrated.

[0079] The processes, methods, or algorithms may be embodied, in whole or in part, using suitable hardware components such as application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), state machines, control units, or other hardware components or devices, or a combination of hardware, software, and firmware components.

[0080] Fig. 6 illustrates a schematic diagram of a control system 502 configured to control a vehicle 600, which may be an at least partially autonomous vehicle or an at least partially autonomous robot. The vehicle 600 includes an actuator 504 and a sensor 506. The sensor 506 may include one or more video sensors, cameras, radar sensors, ultrasonic sensors, LiDAR sensors, and / or position sensors (e.g., GPS). One or more of the one or more particular sensors may be integrated into the vehicle 600. Alternatively, or in addition to one or more particular sensors identified above, the sensor 506 may include a software module configured to determine a state of an actuator 504 based on an execution.A non-limiting example of a software module includes a weather information software module configured to determine a current or future state of the weather in the vicinity of the vehicle 600 or at a further location.

[0081] The classifier 514 of a control system 502 of a vehicle 600 may be configured to detect objects in the vicinity of a vehicle 600 based on input signals x. In such an embodiment, the output signal y may contain information indicating the proximity of objects to the vehicle 600. The actuator control instruction 510 may be determined according to this information. The actuator control instruction 510 may be used to avoid collisions with the detected objects.

[0082] In certain embodiments, the vehicle 600 is an at least partially autonomous vehicle, and the actuator 504 may be embodied in a brake, propulsion system, engine, drivetrain, or steering system of a vehicle 600. The actuator control instructions 510 may be determined such that the actuator 504 is controlled such that the vehicle 600 avoids collisions with detected objects. Detected objects may also be classified according to which classifier 514 considers them most likely to be, for example, pedestrians or trees. The actuator control instructions 510 may be determined depending on the classification. In a scenario where a hostile attack may occur, the system described above may be further trained to better detect objects or to identify a change in lighting conditions or angle for a sensor or camera on the vehicle 600.

[0083] In certain embodiments, where the vehicle 600 is an at least partially autonomous robot, the vehicle 600 may be a mobile robot configured to perform one or more functions such as flying, swimming, diving, and walking. The mobile robot may be an at least partially autonomous lawnmower or an at least partially autonomous cleaning robot. In such embodiments, the actuator control instruction 510 may be determined such that a propulsion unit, a steering unit, and / or a braking unit of the mobile robot may be controlled such that the mobile robot can avoid collisions with identified objects.

[0084] In certain embodiments, the vehicle 600 is an at least partially autonomous robot in the form of a gardening robot. In such an embodiment, the vehicle 600 may use an optical sensor as sensor 506 to determine a condition of plants in an environment near the vehicle 600. The actuator 504 may be a nozzle configured to spray chemicals. Depending on an identified variety and / or an identified condition of the plants, the actuator control instruction 510 may be determined such that an actuator 504 is caused to spray the plants with an appropriate amount of suitable chemicals.

[0085] The vehicle 600 may be an at least partially autonomous robot in the form of a household appliance. Non-limiting examples of household appliances include a washing machine, a stove, an oven, a microwave, or a dishwasher. In such a vehicle 600, the sensor 506 may be an optical sensor configured to detect a condition of an object to be processed by the household appliance. For example, if the household appliance is a washing machine, the sensor 506 may detect a condition of the laundry in the washing machine. The actuator control instruction 510 may be determined based on the detected condition of the laundry.

[0086] Fig. 7 illustrates a schematic diagram of a control system 502 configured to control a system 700 (e.g., a manufacturing machine), such as a punch, cutter, or gun drill of the manufacturing system 702, such as as part of a production line. The control system 502 may be configured to control an actuator 504 configured to control a system 700 (e.g., the manufacturing machine).

[0087] The sensor 506 of the system 700 (e.g., the manufacturing machine) may be an optical sensor configured to capture one or more characteristics of a manufactured product 704. The classifier 514 may be configured to determine a state of a manufactured product 704 from one or more of the captured characteristics. The actuator 504 may be configured to control the system 700 (e.g., the manufacturing machine) for a subsequent manufacturing step of a manufactured product 704 depending on the determined state of a manufactured product 704. The actuator 504 may be configured to control functions of the system 700 (e.g., the manufacturing machine) on a subsequently manufactured product 706 of the system 700 (e.g., the manufacturing machine) depending on the determined state of the manufactured product 704.

[0088] Fig. 8 illustrates a schematic diagram of a control system 502 configured to control a power tool 800, such as a drill or screwdriver, having an at least partially autonomous mode. The control system 502 may be configured to control an actuator 504 configured to control the power tool 800.

[0089] The sensor 506 of the power tool 800 may be an optical sensor configured to record one or more characteristics of a work surface 802 and / or a fastener 804 driven into the work surface 802. The classifier 514 may be configured to determine a condition of a work surface 802 and / or a fastener 804 relative to the work surface 802 from one or more of the recorded characteristics. The condition may be that a fastener 804 is flush with the work surface 802. Alternatively, the condition may be a hardness of a work surface 802. The actuator 504 may be configured to control a power tool 800 such that the drive function of the power tool 800 is adjusted depending on the determined condition of the fastener 804 relative to the work surface 802 or one or more recorded characteristics of the work surface 802.For example, actuator 504 may terminate the drive function if the state of fastener 804 is flush with work surface 802. As another non-limiting example, actuator 504 may apply additional or lesser torque depending on the hardness of a work surface 802.

[0090] Fig. 9 illustrates a schematic diagram of a control system 502 configured to control an automated personal assistant 900. The control system 502 may be configured to control an actuator 504 configured to control an automated personal assistant 900. The automated personal assistant 900 may be configured to control a household appliance such as a washing machine, a stove, an oven, a microwave, or a dishwasher.

[0091] Sensor 506 may be an optical sensor and / or an audio sensor. The optical sensor may be configured to receive video images of gestures 904 from user 902. The audio sensor may be configured to receive a voice command from user 902.

[0092] The control system 502 of an automated personal assistant 900 may be configured to determine actuator control instructions 510 configured in the control system 502. The control system 502 may be configured to determine actuator control instructions 510 according to sensor signals 508 from the sensor 506. The automated personal assistant 900 is configured to send sensor signals 508 to the control system 502. The classifier 514 of the control system 502 may be configured to execute a gesture recognition algorithm to identify a gesture 904 made by the user 902, determine actuator control instructions 510, and send the actuator control instructions 510 to the actuator 504. The classifier 514 may be configured to retrieve information from the non-volatile memory in response to the gesture 904 and output the retrieved information in a form suitable for receipt by the user 902.

[0093] Fig. 10 illustrates a schematic diagram of a control system 502 configured to control a control system 1000. The control system 1000 may be configured to physically control access through the door 1002. The sensor 506 may be configured to detect a scene relevant in deciding whether to grant access. The sensor 506 may be an optical sensor configured to generate and transmit image and / or video data. Such data may be used by the control system 502 to detect a person's face.

[0094] The classifier 514 of the control system 502 of the control system 1000 may be configured to interpret the image and / or video data by comparing identities of known individuals stored in a non-volatile memory 516, thereby determining the identity of an individual. The classifier 514 may be configured to generate an actuator control instruction 510 in response to the interpretation of the image and / or video data. The control system 502 is configured to send the actuator control instruction 510 to the actuator 504. In this embodiment, the actuator 504 may be configured to lock or unlock the door 1002 in response to the actuator control instruction 510. In certain embodiments, non-physical, logical access control is also possible.

[0095] The control system 1000 may also be a surveillance system. In such an embodiment, the sensor 506 may be an optical sensor configured to detect a scene under surveillance, and the control system 502 is configured to control a display device 1004. The classifier 514 is configured to determine a classification of a scene, e.g., whether the scene detected by the sensor 506 is suspicious. The control system 502 is configured to send an actuator control instruction 510 to the display device 1004 in response to the classification. The display device 1004 may be configured to adjust the displayed content in response to the actuator control instruction 510. For example, the display device 1004 may highlight an object deemed suspicious by the classifier 514.Using an embodiment of the disclosed system, the monitoring system can predict that objects will appear at certain times in the future.

[0096] Fig.11 illustrates a schematic diagram of a control system 502 configured to control an imaging system 1100, e.g., an MRI device, an x-ray imaging device, or an ultrasound device. Sensor 506 may be, for example, an imaging sensor. Classifier 514 may be configured to determine a classification of all or a portion of the acquired image. Classifier 514 may be configured to determine or select an actuator control instruction 510 in response to the classification obtained by the trained neural network. For example, classifier 514 may interpret a region of an acquired image as potentially anomalous. In this case, actuator control instruction 510 may be determined or selected to cause display device 1102 to display the image and highlight the potentially anomalous region.

[0097] While exemplary embodiments are described above, these embodiments are not intended to describe all possible forms encompassed by the claims. The words used in the application text are words of description rather than limitation, and it is understood that various changes may be made without departing from the spirit and scope of the disclosure. As previously described, the features of various embodiments may be combined to form further embodiments of the invention, which need not be expressly described or illustrated.While various embodiments may have been described as providing advantages or being preferred over other prior art embodiments or implementations with respect to one or more desired characteristics, those skilled in the art will recognize that one or more features or characteristics may be compromised to achieve desired overall system characteristics depending on the specific application and implementation. These features may include, but are not limited to, cost, strength, durability, life cycle cost, marketability, appearance, packaging, size, usability, weight, manufacturability, ease of assembly, etc.Therefore, to the extent that any embodiments are described as less desirable than other embodiments or prior art implementations with respect to one or more characteristics, those embodiments are not outside the scope of the disclosure and may be desirable for certain applications.

Claims

[1] A method for performing object detection in the open world, the method comprising: Receiving object data, wherein the object data includes embedding data corresponding to embeddings for known objects in a first input image; Projecting the embeddings into a hyperbolic embedding space, wherein the hyperbolic embedding space contains embeddings in a plurality of categories of objects and each of the plurality of categories of objects contains one or more classes of objects; Regularizing the projected embeddings in the hyperbolic embedding space, wherein regularizing the projected embeddings in the hyperbolic embedding space comprises moving each of the projected embeddings (i) closer to embeddings in the same category of the plurality of categories and (ii) farther from embeddings in different categories of the plurality of categories; Receiving an unmatched query corresponding to an object in a second input image; and Generating, based on the hyperbolic embedding space containing the regularized embeddings, an output signal indicating whether the object in the second input image corresponds to an unknown object in one of the classes of objects. [2] The method of claim 1, wherein the object data includes a set of bounding boxes and class labels for the known objects in the first input image. [3] The method of claim 1, wherein the first input image corresponds to a training image. [4] The method of claim 1, further comprising determining a hyperbolic contrastive loss corresponding to the hyperbolic embedding space, wherein regularizing the projected embeddings comprises regularizing the projected embeddings based on the hyperbolic contrastive loss. [5] The method of claim 1, wherein regularizing the projected embeddings comprises determining respective hyperbolic averages of embeddings in each of the classes of objects. [6] The method of claim 5, wherein generating the output signal comprises (i) determining a threshold distance based on the hyperbolic averages and (ii) generating the output signal based in part on the threshold distance. [7] The method of claim 6, wherein generating the output signal comprises determining whether the unmatched sample is less than the threshold distance from at least one of the hyperbolic averages. [8] A computing device configured to perform object detection in the open world, the computing device including a processing device configured to execute instructions stored in a memory to: Receiving object data, wherein the object data includes embedding data corresponding to embeddings for known objects in a first input image; Projecting the embeddings into a hyperbolic embedding space, wherein the hyperbolic embedding space contains embeddings in a plurality of categories of objects and each of the plurality of categories of objects contains one or more classes of objects; Regularizing the projected embeddings in the hyperbolic embedding space, wherein regularizing the projected embeddings in the hyperbolic embedding space comprises moving each of the projected embeddings (i) closer to embeddings in the same category of the plurality of categories and (ii) farther from embeddings in different categories of the plurality of categories; Receiving an unmatched query corresponding to an object in a second input image; and Generating, based on the hyperbolic embedding space containing the regularized embeddings, an output signal indicating whether the object in the second input image corresponds to an unknown object in one of the classes of objects. [9] The computing device of claim 8, wherein the object data includes a set of bounding boxes and class labels for the known objects in the first input image. [10] A computing device according to claim 8, wherein the first input image corresponds to a training image. [11] The computing device of claim 8, wherein the processing device is further configured to execute instructions stored in a memory to generate a hyperbolic contrastive loss corresponding to the hyperbolic embedding space, wherein regularizing the projected embeddings comprises regularizing the projected embeddings based on the hyperbolic contrastive loss. [12] The computing device of claim 8, wherein the processing device for regularizing the projected embeddings is further configured to execute instructions stored in a memory to determine respective hyperbolic averages of embeddings in each of the classes of objects. [13] The computing device of claim 12, wherein the processing device for generating the output signal is further configured to execute instructions stored in a memory for (i) determining a threshold distance based on the hyperbolic averages and (ii) generating the output signal based in part on the threshold distance. [14] The computing device of claim 13, wherein the processing device for generating the output signal is further configured to execute instructions stored in a memory to determine whether the unmatched sample is less than the threshold distance from at least one of the hyperbolic averages. [15] Computer-controlled machine comprising: at least one sensor configured to generate a first input image and a second input image; a control system configured to: Receiving object data, wherein the object data includes embedding data corresponding to embeddings for known objects in a first input image, Projecting the embeddings into a hyperbolic embedding space, wherein the hyperbolic embedding space contains embeddings in a plurality of categories of objects and each of the plurality of categories of objects contains one or more classes of objects, Regularizing the projected embeddings in the hyperbolic embedding space, wherein regularizing the projected embeddings in the hyperbolic embedding space comprises moving each of the projected embeddings (i) closer to embeddings in the same category of the plurality of categories and (ii) further away from embeddings in different categories of the plurality of categories, Receiving an unmatched query corresponding to an object in a second input image, and Generating, based on the hyperbolic embedding space containing the regularized embeddings, an output signal indicating whether the object in the second input image corresponds to an unknown object in one of the classes of objects; and an actuator configured to control operation of the computer-controlled machine in response to the output signal. [16] A computer-controlled machine according to claim 15, further comprising a memory that stores data corresponding to the hyperbolic embedding space. [17] The computer-controlled machine of claim 15, wherein the control system is further configured to generate a hyperbolic contrastive loss corresponding to the hyperbolic embedding space, wherein regularizing the projected embeddings comprises regularizing the projected embeddings based on the hyperbolic contrastive loss. [18] The computer-controlled machine of claim 15, wherein the control system for regularizing the embeddings is further configured to determine respective hyperbolic averages of embeddings in each of the classes of objects. [19] The computer-controlled machine of claim 18, wherein the control system for generating the output signal is further configured to (i) determine a threshold distance based on the hyperbolic averages and (ii) determine whether the unmatched sample is less than the threshold distance of at least one of the hyperbolic averages. [20] A computer-controlled machine according to claim 15, wherein the computer-controlled machine includes an autonomous robot.

Images