Abstract: The present invention extends to methods, systems, and computer program products for profiling video output devices. A frameset is received from a video output device. A first characteristic of the frameset is profiled. A focus area within the frameset is identified based on the profile of the first characteristic. A second characteristic is profiled within the focus area. A baseline output profile for the video device is generated based at least on the profile of the second characteristic. The baseline output profile is stored in repository.

Abstract: The present invention extends to methods, systems, and computer program products for classifying video. In one aspect, a model is trained to classify video. A reference model capable of identifying signals from a plurality of signals classes is used along with tagged training data to train a signal class specific model. The signal class specific model uses less resources to identify the specific signal class model relative to the reference model. The signal class specific model can be deployed to an event detection infrastructure or other device. Frames from video streams can be classified using the signal class specific model.

Abstract: Examples extend to methods, systems, and computer program products for disrupting object recognition functionality. Images can be altered in a manner that makes it difficult (if not impossible) for recognition algorithms to accurately recognize faces, vehicles, people, etc. in the images. Image alterations can be tailored to disrupt recognition algorithms while being imperceptible to the human eye and/or minimizing impact on other image processing systems. In some aspects, Machine Leaning (ML) frameworks and/or leaning-based generative/discriminative model techniques, such as, for example, Generative Adversarial Networks (GANs), are used to alter images.

Abstract: The present invention extends to methods, systems, and computer program products for detecting and validating planned event information. The dimensionality of a raw signal is reduced into a normalized signal that includes a probability value. The probability value at least approximates a probability that the normalized signal indicates that a real-world event of a specified event type is occurring. Based on the at least approximated probability, the real-world event of the specified event type is detected as a possible planned event occurring at a venue. The possible planned event is validated as an actual planned event. Validation can include validating data associated with the possible planned event with data from an event planning system associated with the venue. An entity is electronically notified that the actual planned event is occurring.

Abstract: Examples extend to methods, systems, and computer program products for managing roadway incidents. A probable origination location of a roadway incident is identified from features of a normalized signal. One or more additional normalized signals within a specified distance of the probable origination location are accessed. The probable origination location is validated, from features of the one or more additional signals, to establish a validated origination location. An event associated with the roadway incident is detected from the features of the normalized signal based on the validated origination location. The detected event includes the validated origination location and a probability that the event is true. Dispatch of resources responding to the roadway incident event is tailored based on the validated origination location and the calculated probability.

Abstract: The present invention extends to methods, systems, and computer program products for filtering signals during major events. A major event is detected in a geographic area based on one or more of: signal volume, signal diversity, severity, content, or historical events associated with ingested digital signals corresponding to the geographic area. An event-specific filter is deployed. The region associated with the major event is locked to the geographic area. Filtering out a commentary signal purportedly related to the major event is filtered out in accordance with rejection criteria. Filtering out the commentary signal can include determining the commentary signal originated outside the geographic area. It is determined that the major event has ended. The event-specific filter is disabled.

Abstract: The present invention extends to methods, systems, and computer program products for detecting major events. A sequence of normalized signals is accessed. An event is detected at a location based on the signal sequence. An interested party is notified of the event detection. The detected event is compared to (historical and/or current) major event data associated with the location. It is determined that the detected event is anomaly at the location in view of results of the comparison. The event is classified as a major event based on the detected event being an anomaly. Subsequent events outside a buffer around the major event can be identified as human ripple effects.

Abstract: The present invention extends to methods, systems, and computer program products for creating signal sequences. Signals in a signal sequence that lack sufficient similarity to other signals in the signal sequence can be moved to a different signal sequence. Similarity can be computed based on signal contexts as well as the (location and/or time) distance between signals.

Abstract: The present invention extends to methods, systems, and computer program products for segmenting video stream frames. In one aspect, video stream frames are segmented into more relevant segments (e.g., including roadway) and less relevant segments (e.g., not including roadway). Different segments can be handled differently. For example, more relevant segments can be processed to identify vehicles, identify events, etc. and less relevant segments may be ignored. Accordingly, resources can be utilized more efficiently. In one aspect, a binary mask is generated from object data in one or more frames. The binary mask is applied to further frames blocking out less relevant frame segments in the further frames.

Abstract: The present invention extends to methods, systems, and computer program products for assessing video stream quality. The quality of a video stream is classified into one of a plurality of quality classifications including a low quality classification and at least one other higher quality classification. A plurality of video quality thresholds are accessed. If any one of the plurality of video quality thresholds is not satisfied, the video stream is classified as low quality. The characteristics of a video stream frame are computed to satisfy each of the plurality of quality thresholds. A video stream technical score is computed from content of the frame based on and subsequent to satisfaction of the plurality of quality thresholds. The video stream is classified as a specified quality, from among the plurality of quality classifications, based on the video stream technical score.

Abstract: The present invention extends to methods, systems, and computer program products for estimating color of vehicles on a roadway. The color can be at least estimated from captured imagery, such as, frames of a video stream. The captured imagery can be captured by variably configured traffic cameras, other public cameras, private cameras (e.g., CCTV), or virtually any other camera, etc., which may have different operating conditions. Color confidence scores can be derived based on image characteristics.

Abstract: The present invention extends to methods, systems, and computer program products for detecting and validating planned event information. The dimensionality of a raw signal is reduced into a normalized signal that includes a probability value. The probability value at least approximates a probability that the normalized signal indicates that a real-world event of a specified event type is occurring. Based on the at least approximated probability, the real-world event of the specified event type is detected as a possible planned event occurring at a venue. The possible planned event is validated as an actual planned event. Validation can include validating data associated with the possible planned event with data from an event planning system associated with the venue. An entity is electronically notified that the actual planned event is occurring.

Abstract: The present invention extends to methods, systems, and computer program products for classifying video. In one aspect, a model is trained to classify video. A reference model capable of identifying signals from a plurality of signals classes is used along with tagged training data to train a signal class specific model. The signal class specific model uses less resources to identify the specific signal class model relative to the reference model. The signal class specific model can be deployed to an event detection infrastructure or other device. Frames from video streams can be classified using the signal class specific model.

Abstract: The present invention extends to methods, systems, and computer program products for estimating color of vehicles on a roadway. The color can be at least estimated from captured imagery, such as, frames of a video stream. The captured imagery can be captured by variably configured traffic cameras, other public cameras, private cameras (e.g., CCTV), or virtually any other camera, etc., which may have different operating conditions.

Abstract: The present invention extends to methods, systems, and computer program products for estimating color of vehicles on a roadway. The color can be at least estimated from captured imagery, such as, frames of a video stream. The captured imagery can be captured by variably configured traffic cameras, other public cameras, private cameras (e.g., CCTV), or virtually any other camera, etc., which may have different operating conditions. Color confidence scores can be derived based on image characteristics.

Abstract: The invention extends to methods, systems, and computer program products for detecting events from a signal features matrix. A two-dimensional signal evidence matrix is formed is populated with the plurality of signal pairings. Per signal pairing, a pairing probability is computed based on one or more of: (a) source diversity between the signals in the signal pairing, (b) pairing frequency indicating how often signal types corresponding to the signals in the signal pair are paired together, (c) pairing strength derived from a confidence associated with each signal in the signal pairing, (d) pairing time derived from a time associated with each signal in the signal pairing, or (e) pairing location derived from a location associated with each signal in the signal pairing. The plurality of pairing probabilities is aggregated into an aggregated probability. The real-world event is detected from evidence provided by the aggregated probability.

Abstract: The present invention extends to methods, systems, and computer program products for detecting and validating planned event information. The dimensionality of a raw signal is reduced into a normalized signal that includes a probability value. The probability value at least approximates a probability that the normalized signal indicates that a real-world event of a specified event type is occurring. Based on the at least approximated probability, the real-world event of the specified event type is detected as a possible planned event occurring at a venue. The possible planned event is validated as an actual planned event. Validation can include validating data associated with the possible planned event with data from an event planning system associated with the venue. An entity is electronically notified that the actual planned event is occurring.

Abstract: The present invention extends to systems and methods for notifying entities of relevant events. An entity defines a rule formula that is triggered when one or more detected events match the rule formula including defining one or more event types and one or more locations types. A boundary associated with a selected location type is also received along with a monitoring area. The received elements are combined into the rule formula notification preferences are associated with the defined rule. One or more events are detected and then compared to the rule formula to determine if the combination of one or more event types occurred within the boundary of the one or more location types within the area.

Abstract: The present invention extends to methods, systems, and computer program products for detecting and validating planned event information. A plurality of normalized signals is accessed. Planned event data across the plurality of normalized signals is checked for inconsistencies. Any inconsistencies are resolved in an automated fashion, for example, through reference to databases containing additional information. A planned event can be detected/validated from concurring and/or resolved planned event data. A validator can refer to an event history database and/or a planning system to validate a possible planned event as an actual planned event.

Abstract: The present invention extends to methods, systems, and computer program products for detecting and validating planned event information. A plurality of normalized signals is accessed. Planned event data across the plurality of normalized signals is checked for inconsistencies. Any inconsistencies are resolved in an automated fashion, for example, through reference to databases containing additional information. A planned event can be detected/validated from concurring and/or resolved planned event data. A validator can refer to an event history database and/or a planning system to validate a possible planned event as an actual planned event.