CN116205277A - Event processing method and device, electronic device, computer readable medium - Google Patents

Abstract

The present disclosure provides an event processing method and apparatus, an electronic device, and a computer readable medium, where the method includes: acquiring an event stream to be processed; performing event detection on the event stream, and determining the current event detection result of the event stream; according to the event detection result, adjusting the time step for processing the event stream; and processing the event stream by adopting the time step to obtain a processing result of the event stream. According to the embodiment of the disclosure, the operation amount in the processing process can be reduced on the basis of ensuring the processing accuracy.

Description

Event processing method and device, electronic device, computer readable medium

technical field

The present disclosure relates to the field of computer technology, and in particular to an event processing method and device, electronic equipment, and a computer-readable medium.

Background technique

Neuromorphic technology refers to a technology that uses electronic technology to simulate the structure of the nervous system in organisms for data processing. In related technologies, a neuromorphic model (also known as a spiking neural network) can be implemented through a hardware circuit (such as a neuromorphic chip) or a software system.

When the system uses a neuromorphic model for processing (such as brain simulation), it is usually necessary to pre-set the time step before running, and use this time step as the cycle for processing. The time step is usually a fixed value, such as 0.1 ms, that is, for example, 10,000 biological time beats are processed per second in the brain simulation process.

In the related technology, the time step needs to be set small enough to process the required detail changes, which will lead to an excessive amount of calculation; and if a larger step is set, the details may be lost, and even the accuracy of the processing cannot be guaranteed. sex and authenticity.

Contents of the invention

The present disclosure provides a many-core system-based event processing method and device, electronic equipment, and a computer-readable medium.

In a first aspect, the present disclosure provides an event processing method, which includes:

Obtain the event stream to be processed; perform event detection on the event stream, and determine the current event detection result of the event stream; adjust the time step for processing the event stream according to the event detection result; adopt the The time step processes the event flow to obtain a processing result of the event flow.

In some possible implementation manners, the event detection result includes information about whether an event is detected in the event flow, wherein the adjusting the time step for processing the event flow according to the event detection result includes : When event information is detected in the event stream, adjust the time step to a second time step, where the second time step is smaller than the initial first time step.

In some possible implementation manners, the event detection result includes information about whether an event is detected in the event flow, wherein the adjusting the time step for processing the event flow according to the event detection result includes : When no event information is detected in the event flow, and the current time step is the second time step, adjust the time step to the initial first time step, and the second time step is shorter than the first time step.

In some possible implementation manners, the event stream includes multiple areas, and the event detection result includes whether an event is detected in each area of the event stream, wherein, according to the event detection result, adjusting the The time step in which the event stream is processed, including:

If there is a target area in the plurality of areas, adjust the time step of the target area to a second time step, the target area is an area where event information is detected, and the second time step smaller than the initial first time step.

In some possible implementation manners, the event detection result includes information about whether an event is detected in the event flow, and the event density of the event flow, wherein, according to the event detection result, the adjustment to the event A time step for stream processing, including: adjusting the time step according to the event density when event information is detected in the event stream,

Wherein, adjusting the time step according to the event density includes: when the event density is greater than or equal to a first density threshold and less than or equal to a second density threshold, adjusting the time step to the second density threshold Two time steps, the second density threshold is greater than the first density threshold, the second time step is smaller than the initial first time step; or in the case where the event density is greater than the second density threshold, The time step is adjusted to a third time step, the third time step being smaller than the second time step.

In some possible implementation manners, the performing event detection on the event flow and determining the current event detection result of the event flow includes: performing event detection on the event flow and determining the current event density of the event flow ; In a case where the event density is greater than or equal to a third density threshold, determining that the event detection result is: event information is detected in the event flow.

In some possible implementation manners, the performing event detection on the event flow and determining the current event detection result of the event flow includes: dividing the event flow into multiple event frames according to a preset duration; An event detection is performed on the current event frame of the event flow through an event detection network to obtain the event detection result.

In some possible implementation manners, the processing the event flow by using the time step to obtain the processing result of the event flow includes: using the time step to process the event through a spiking neural network The flow is subjected to brain simulation processing to obtain the response information of the spiking neural network for the event flow, and the processing result includes the response information, wherein the differential expression of the neuron parameters in the brain simulation is performed according to the time step Adjustment.

In some possible implementation manners, the event stream is an event stream of a target scene collected by an event collection device, and the event stream is used to represent brightness change information of the target scene.

In a second aspect, the present disclosure provides an event processing device, which includes:

The event flow obtaining module is used to obtain the event flow to be processed; the event detection module is used to perform event detection on the event flow and determine the current event detection result of the event flow; the step adjustment module is used to The event detection result is used to adjust the time step for processing the event flow; the event flow processing module is used to process the event flow by using the time step to obtain the processing result of the event flow.

In some possible implementation manners, the event detection result includes whether event information is detected in the event stream, wherein the step adjustment module is configured to: if event information is detected in the event stream , adjusting the time step to a second time step, the second time step being smaller than the initial first time step.

In some possible implementation manners, the event detection result includes whether event information is detected in the event stream, wherein the step adjustment module is configured to: no event information is detected in the event stream, and When the current time step is the second time step, the time step is adjusted to the initial first time step, and the second time step is smaller than the first time step.

In some possible implementation manners, the event stream includes multiple areas, and the event detection result includes whether an event is detected in each area of the event stream, wherein the step adjustment module is configured to: When there is a target area in the plurality of areas, adjust the time step of the target area to a second time step, the target area is an area where event information is detected, and the second time step is less than The initial first time step.

In some possible implementation manners, the event detection result includes whether event information is detected in the event flow, and the event density of the event flow, wherein the step adjustment module includes: an adjustment submodule configured to in case event information is detected in the event stream, adjusting the time step according to the event density,

Wherein, the adjustment submodule is configured to: adjust the time step to a second time step when the event density is greater than or equal to a first density threshold and less than or equal to a second density threshold, The second density threshold is greater than the first density threshold, and the second time step is smaller than the initial first time step; or if the event density is greater than the second density threshold, the time step is The length is adjusted to a third time step, and the third time step is smaller than the second time step.

In some possible implementations, the event detection module is configured to: perform event detection on the event flow, and determine the current event density of the event flow; when the event density is greater than or equal to a third density threshold Next, it is determined that the event detection result is: event information is detected in the event flow.

In some possible implementations, the event detection module is configured to: divide the event stream into multiple event frames according to a preset duration; Detect to obtain the event detection result.

In some possible implementation manners, the event flow processing module is configured to: use the time step to perform brain simulation processing on the event flow through the spiking neural network, and obtain the The response information of the processing result includes the response information, wherein the differential expressions of the neuron parameters in the brain simulation are adjusted according to the time step.

In some possible implementation manners, the event stream is an event stream of a target scene collected by an event collection device, and the event stream is used to represent brightness change information of the target scene.

In a third aspect, the present disclosure provides an electronic device, which includes: at least one processor; and a memory communicated with the at least one processor; wherein, the memory stores information that can be processed by the at least one processor. One or more computer programs executed by the processor, and one or more computer programs are executed by the at least one processor, so that the at least one processor can execute the above event processing method.

In a fourth aspect, the present disclosure provides a computer-readable medium on which a computer program is stored, wherein the computer program implements the above-mentioned event processing method when executed by a processor.

The embodiment provided by the present disclosure can perform event detection on the event flow to determine the event detection result; adjust the time step for processing the event flow according to the event detection result; use the time step to process the event flow to obtain the processed As a result, the dynamic adjustment of the time step is realized, and the calculation amount in the whole processing process is reduced on the basis of ensuring the accuracy and authenticity of the processing.

It should be understood that what is described in this section is not intended to identify key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will be readily understood through the following description.

Description of drawings

The accompanying drawings are used to provide a further understanding of the present disclosure, and constitute a part of the specification, and are used together with the embodiments of the present disclosure to explain the present disclosure, and do not constitute a limitation to the present disclosure. The above and other features and advantages will become more apparent to those skilled in the art by describing detailed example embodiments with reference to the accompanying drawings, in which:

FIG. 1 is a flowchart of an event processing method provided by an embodiment of the present disclosure;

FIG. 2 is a schematic diagram of a processing process of an event processing method provided by an embodiment of the present disclosure;

FIG. 3 is a block diagram of an event processing device provided by an embodiment of the present disclosure;

Fig. 4 is a block diagram of an electronic device provided by an embodiment of the present disclosure.

Detailed ways

In order for those skilled in the art to better understand the technical solution of the present disclosure, the exemplary embodiments of the present disclosure are described below in conjunction with the accompanying drawings, which include various details of the embodiments of the present disclosure to facilitate understanding, and they should be considered exemplary only. Accordingly, those of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the disclosure. Also, descriptions of well-known functions and constructions are omitted in the following description for clarity and conciseness.

In the case of no conflict, various embodiments of the present disclosure and various features in the embodiments can be combined with each other.

As used herein, the term "and/or" includes any and all combinations of one or more of the associated listed items.

The terminology used herein is for describing particular embodiments only and is not intended to limit the present disclosure. As used herein, the singular forms "a" and "the" are intended to include the plural forms as well, unless the context clearly dictates otherwise. It will also be understood that when the terms "comprising" and/or "consisting of" are used in this specification, the stated features, integers, steps, operations, elements and/or components are specified to be present but not excluded to be present or Add one or more other features, integers, steps, operations, elements, components and/or groups thereof. Words such as "connected" or "connected" are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect.

Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art. It will also be understood that terms such as those defined in commonly used dictionaries should be interpreted as having meanings consistent with their meanings in the context of the relevant art and the present disclosure, and will not be interpreted as having idealized or excessive formal meanings, Unless expressly so limited herein.

As mentioned above, in related technologies, when neuromorphic models are used for processing (such as brain simulation), a fixed time step is preset before running, which may lead to excessive calculation or loss of details.

However, the neurons in the neuromorphic model are driven by events, that is, the neurons start computing when there is an input event, and the system usually pays more attention to the response of the neuron after receiving the input (stimulus).

According to the event processing method of the embodiment of the present disclosure, the time step can be dynamically adjusted during the processing of input information (such as event flow), and when a stimulus signal is received, a finer time step can be adopted to improve the processing accuracy; When there is no input or no stimulus signal (also known as the signal of interest), a coarser time step is used to reduce the amount of calculation of the system, thereby reducing the processing time while ensuring the accuracy and authenticity of the processing. The amount of calculations during the entire processing.

The event processing method according to the embodiments of the present disclosure may be executed by electronic devices such as terminal devices or servers, and the terminal devices may be vehicle-mounted devices, user equipment (User Equipment, UE), mobile devices, user terminals, terminals, cellular phones, cordless phones, For a personal digital assistant (Personal Digital Assistant, PDA), a handheld device, a computing device, a vehicle-mounted device, a wearable device, etc., the method can be realized by calling a computer-readable instruction stored in a memory by a processor. Alternatively, the method may be performed by a server.

FIG. 1 is a flowchart of an event processing method provided by an embodiment of the present disclosure. Referring to Figure 1, the method includes:

In step S11, obtain the event flow to be processed;

In step S12, event detection is performed on the event flow, and the current event detection result of the event flow is determined;

In step S13, adjusting the time step for processing the event flow according to the event detection result;

In step S14, the event flow is processed by using the time step to obtain a processing result of the event flow.

For example, the event stream to be processed may be obtained in step S11, and the event stream may be a pulse signal (or called a pulse train) with spatial information and time information. For example, the event flow corresponds to M*N spatial positions (M, N are integers greater than 1), and each spatial position has a pulse signal generated with time t. The pulse signal can be a binary sequence of 0,1, or a three-valued sequence of -1,0,1, and the present disclosure does not limit the specific expression form of the pulse signal.

In some possible implementation manners, the event stream may be an event stream of a target scene collected by an event collection device, and the event stream is used to represent brightness change information of the target scene.

Wherein, the target scene may be a scene including geographic areas corresponding to buildings, scenery, people, vehicles, and the like. The event collection device can be, for example, an event camera (Event Camera), which can collect asynchronous brightness changes (ie events (Event)) with a high temporal frequency, and output event data in stream form (event stream).

In some possible implementation manners, the initial event flow collected by the event collection device may be binarized or trivaluated to obtain the event flow to be processed, and the disclosure does not limit the specific processing manner.

Those skilled in the art should understand that the event stream to be processed may also be any event data in stream form with other physical meanings, such as voice, text, etc., and the disclosure does not limit the specific content of the event stream.

In some possible implementation manners, in step S12, event detection may be performed on the event stream, and a current event detection result of the event stream may be determined. The event detection result includes whether event information is detected in the event flow, that is, event information is detected or no event information is detected. The present disclosure does not limit the specific manner of event detection.

In some possible implementations, the density of event information in the event stream can be counted in real time, for example, among the M*N spatial positions of the current event stream, the positions with event information (for example, a value of 1) are in all spaces The proportion in the position is used as the event density; when the event density is greater than or equal to the preset density threshold, it is determined that the event information is detected; otherwise, when the event density is less than the density threshold, it is determined that the event information is not detected.

In some possible implementation manners, the corresponding event detection network may also be pre-trained, and the current event flow is input to the event detection network for classification to determine whether event information is detected. The event detection network is, for example, a convolutional neural network or a spiking neural network, which is not limited in the present disclosure.

In some possible implementation manners, in step S13, the time step for the system to process the event flow may be adjusted according to the event detection result. For example, you can use a larger time step at the beginning of the run; adjust to a smaller time step after the event information is detected; adjust back to a larger time step after the event information ends; The time steps of multiple gears can be set according to the quantity/density of event information, etc., and the present disclosure does not limit the specific adjustment method.

In some possible implementation manners, in step S14, the system may process the event flow by using the above-mentioned time step, for example, perform brain simulation to obtain a processing result of the event flow, so as to realize corresponding processing tasks.

FIG. 2 is a schematic diagram of a processing process of an event processing method provided by an embodiment of the present disclosure. Referring to FIG. 2 , an event detection component 21 and an event stream processing component 22 may be provided for event detection and event stream processing respectively. The event detection component 21 and the event flow processing component 22 may be hardware modules or software modules, which are not limited in this disclosure.

In an example, the input event stream 23 may be input into the event detection component 21 and the event stream processing component 22 respectively. The event detection component 21 performs event detection on the event flow 23 to determine the event detection result; adjusts the time step according to the event detection result; and sends the time step to the event flow processing component 22 . The event flow processing component 22 uses the time step to process the event flow 23, and outputs a corresponding processing result 24, thereby realizing the entire adjustment and processing process.

According to the embodiments of the present disclosure, event detection can be performed on the event flow to determine the event detection result; according to the event detection result, the time step for processing the event flow can be adjusted; the event flow can be processed using the time step to obtain the processing result , so as to realize the dynamic adjustment of the time step, and reduce the calculation amount in the whole processing process on the basis of ensuring the accuracy and authenticity of the processing.

The event processing method according to the embodiment of the present disclosure will be described below.

As mentioned above, the event flow to be processed can be obtained in step S11; and event detection is performed on the event flow in step S12.

In some possible implementations, step S12 may include:

performing event detection on the event flow, and determining the current event density of the event flow;

In a case where the event density is greater than or equal to a third density threshold, it is determined that the event detection result is: event information is detected in the event flow.

For example, event detection can be done statistically. For example, calculate the density of event information in the event stream in real time, and calculate the ratio of the positions with event information (such as 1 and/or -1) in all spatial positions among the M*N spatial positions of the current event stream , as the event density.

In some possible implementations, an event stream of a certain length may also be selected, such as event data within a certain period of time before and after the current time point, and the proportion of event information in the event data is counted as the event density.

In some possible implementations, different weights can also be set for different spatial positions. For example, the weight of the region of interest in the M*N spatial positions is larger, and the weight of other regions is smaller. Events at each spatial position After the information is weighted and summed, the event density of the event flow is determined.

In some possible implementation manners, the M*N spatial locations may also be divided into multiple regions, and the event density of each region is counted separately. The present disclosure does not limit the specific statistical methods.

In some possible implementations, the event information of interest can also be set, for example, the pulse signal with a value of 1 in the event stream is used as the event information of interest, and the pulse signal with a value of -1 is ignored, so that the statistics The event density obtained is the density of event information of interest, which makes event detection more targeted, thereby further improving the effect of event detection.

In some possible implementation manners, if the event density is greater than or equal to a preset density threshold (called a third density threshold), it may be determined that the event detection result is: event information is detected in the event flow. On the contrary, it can be determined that the event detection result is: no event information is detected in the event flow. The present disclosure does not limit the specific value of the third density threshold.

In some possible implementation manners, in the case of counting the event density by region, it may be determined whether event information is detected in each region, that is, the event detection result includes whether event information is detected in each region.

In this way, whether there is event information in the event stream can be detected in a simple manner, which improves the flexibility of detection and reduces the calculation amount of detection.

In some possible implementation manners, performing event detection on the event flow in step S12 may also include:

Dividing the event stream into multiple event frames according to a preset duration;

An event detection is performed on the current event frame of the event flow through an event detection network to obtain the event detection result.

For example, event detection can be performed by means of neural networks. Divide the event stream into multiple event frames according to preset duration. The duration can be set, for example, as the sampling duration of the event stream to improve the real-time performance of the processing; it can also be set as multiple sampling durations to improve the robustness of the processing. The present disclosure does not limit the specific value of the duration.

In some possible implementations, the current event frame of the event stream can be input into the pre-trained event detection network for processing, classify whether the event stream is in the input period, and output the event detection result, that is, whether the event stream detects to event information.

Wherein, the event detection network is, for example, a convolutional neural network or a spiking neural network, and the present disclosure does not limit the specific network structure and training methods of the event detection network.

In this way, the accuracy of event detection can be improved.

In some possible implementation manners, after the event detection result is determined in step S12, the time step for processing the event flow may be adjusted in step S13 according to the event detection result.

In some possible implementations, step S13 may include:

In case event information is detected in the event stream, the time step is adjusted to a second time step, the second time step being smaller than the initial first time step.

That is, if event information is detected in the event stream, the time step for processing the event stream may be adjusted to a smaller time step (called the second time step). The second time step is smaller than the time step when the system is initially running (referred to as the first time step), for example, the first time step is 10 ms, and the second time step is 0.1 ms.

It should be understood that those skilled in the art can set the first time step and the second time step according to the actual situation. The specific values of the first time step and the second time step in this disclosure, as well as the specific The adjustment methods are not limited.

In this way, a finer time step can be used when there is event information, thereby improving the processing accuracy.

In some possible implementations, step S13 may include:

When no event information is detected in the event flow, and the current time step is the second time step, adjust the time step to the initial first time step, and the second time step less than the first time step.

That is, if no event information is detected in the event stream, it can be adjusted according to the current time step. If the current time step is already the larger first time step, no adjustment is made; if the current time step is the smaller second time step, the time step is adjusted back to the larger second time step A time step.

In this way, it is possible to reduce the amount of calculation during the period of no event information.

In some possible implementations, step S13 may include:

If there is a target area in the plurality of areas, adjust the time step of the target area to a second time step, the target area is an area where event information is detected, and the second time step smaller than the initial first time step.

For example, the event stream may include multiple regions, for example, divide M*N spatial positions of the event stream into 9 regions of 3*3. In step S12, it is determined whether event information is detected in each area, that is, the event detection result of the event stream includes whether event information is detected in each area of the event stream. The present disclosure does not limit the specific detection method.

In some possible implementations, the area where the event information is detected can be used as the target area. It is possible that there are moving objects in the target area, such as footballs rolling in the field, vehicles driving on the street, etc.; Objects remain motionless.

In some possible implementations, if there are target regions in multiple regions of the event flow, the time step for processing the target region of the event flow is adjusted to a smaller second time step, and the time for other regions The step size is not adjusted and remains the larger first time step. Conversely, if the target region does not exist in multiple regions, the time step of the entire event stream processing is adjusted to the larger first time step.

In this way, the time step of event stream processing can be adjusted in more detail, the accuracy of step adjustment can be improved, and the calculation amount of the system can be further reduced.

In some possible implementations, the event detection result includes whether event information is detected in the event flow, and the event density of the event flow, where step S13 may include:

in case event information is detected in the event stream, adjusting the time step according to the event density,

Wherein, the step of adjusting the time step according to the event density includes:

When the event density is greater than or equal to a first density threshold and less than or equal to a second density threshold, the time step is adjusted to a second time step, and the second density threshold is greater than the first a density threshold, the second time step being smaller than the initial first time step; or

If the event density is greater than a second density threshold, the time step is adjusted to a third time step, and the third time step is smaller than the second time step.

For example, the event detection result can also include the event density of the event flow, so as to set multiple levels of time step adjustment, so as to perform finer adjustment.

In some possible implementations, if event information is detected in the event stream, the time step can be adjusted according to the event density. Multiple density thresholds for event density may be preset, such as a first density threshold and a second density threshold, and the second density threshold (for example, 0.3) is greater than the first density threshold (for example, 0.1).

In some possible implementations, if the event density is greater than or equal to the first density threshold and less than or equal to the second density threshold, it can be considered that there is low-density event information (there is event information but a small amount), and the time The step size is adjusted to the second time step. The second time step (for example, 1 ms) is smaller than the initial first time step (for example, 10 ms), so that the calculation amount is not increased too much while ensuring the processing accuracy.

In some possible implementation manners, if the event density is greater than the second density threshold, it can be considered that there is high-density event information (there is event information and a large number), and the time step can be adjusted to the third time step. The third time step (eg, 0.1 ms) is smaller than the second time step (eg, 1 ms), so as to improve processing accuracy and authenticity.

In some possible implementation manners, the event stream may also be divided into multiple regions, and the same or different density thresholds and time steps of each gear are set for the multiple regions respectively. For example, set a lower density threshold, a more detailed time step, etc. for the region of interest in the event stream.

It should be understood that those skilled in the art may set the number and value of the density thresholds and the specific value of the time step of each gear according to the actual situation, which is not limited in the present disclosure.

In this way, the accuracy of the step size adjustment can be further improved, and the calculation amount of the system can be further reduced on the basis of ensuring the accuracy and authenticity of the processing.

In some possible implementation manners, in step S14, a corresponding time step is used to process the event flow. Wherein, step S14 may include:

By adopting the time step, brain simulation processing is performed on the event flow through a spiking neural network to obtain response information of the spiking neural network to the event flow, and the processing result includes the response information.

For example, the processing of the event flow can be brain simulation processing, and the processing process of the brain on the event flow is imitated through the spiking neural network. A spiking neural network can be preset, and the spiking neural network includes a plurality of neurons, each neuron is connected to other neurons, and can receive and/or send pulses. The disclosure does not limit the network structure and implementation of the spiking neural network.

In some possible implementation manners, the processing tasks performed by the spiking neural network may include, for example, any one of image processing tasks, speech processing tasks, text processing tasks, video processing tasks, and other tasks. This disclosure does not limit this.

In some possible implementations, the adjusted time step in step S13 can be used to perform brain simulation processing on the event flow through the spiking neural network, and the response information of the spiking neural network to the event flow can be obtained as the processing of the event flow As a result, for example, a moving vehicle in the scene is identified.

In some possible implementations, the differential expressions of the neuron parameters in the brain simulation are adjusted according to the time step. Neuron parameters may, for example, include membrane potential, synaptic delay, etc., and the present disclosure does not limit specific types of neuron parameters.

For example, in the case of variable time steps, various calculation formulas in the brain simulation may change accordingly, where the differential equations (such as dv/dt) in the brain simulation are constant, and the transformation After being expressed as a difference, it is necessary to adjust the variables related to Δt in the differential equation of the system according to the time step Δt, that is, the time step Δt is no longer a constant.

For example, the formula for calculating the membrane potential of a neuron can be expressed as:

Vupd(n)=Vupd(n-1)*φ(Δt)+f(Δt) (1)

In formula (1), Vupd(n) represents the membrane potential of the neuron in the current nth processing cycle; Vupd(n-1) neuron represents the membrane potential of the previous (n-1th) processing cycle means; Changes in membrane potential, and related terms for resting potential.

It can be seen that the change in membrane potential is a function of the time step Δt.

In addition, when the synaptic delay (delay) in a neuron is converted into a differential expression (expressed as the number of time steps), it also needs to be adjusted according to the time step Δt. For example, when the delay is 1ms, if the time step Δt is 0.1ms, the value of the delay is represented as 10; if the time step Δt is 0.01ms, the value of the delay is represented as 100.

In this way, the event stream processing under the dynamic time step can be realized, the adaptability of the processing process can be improved, and the processing accuracy can be improved.

According to the event processing method of the embodiment of the present disclosure, the time step can be dynamically adjusted during the processing of the input event stream, and when event information is detected, a finer time step can be adopted to improve the processing accuracy; When there is no input or event information of interest, a coarser time step is used to reduce the amount of calculation of the system, thereby reducing the amount of calculation in the entire processing process on the basis of ensuring the accuracy and authenticity of the processing.

The event processing method according to the embodiments of the present disclosure can be applied to various application scenarios in various application fields such as brain simulation, computer vision, target detection, and target tracking, so as to realize corresponding processing of event streams and improve computing efficiency.

Fig. 3 is a block diagram of an event processing device provided by an embodiment of the present disclosure.

Referring to FIG. 3 , an embodiment of the present disclosure provides an event processing device, which includes:

An event stream acquiring module 31, configured to acquire event streams to be processed;

An event detection module 32, configured to perform event detection on the event flow, and determine the current event detection result of the event flow;

A step adjustment module 33, configured to adjust the time step for processing the event flow according to the event detection result;

The event flow processing module 34 is configured to process the event flow by using the time step to obtain a processing result of the event flow.

In some possible implementation manners, the event detection result includes whether event information is detected in the event stream, wherein the step adjustment module is configured to: if event information is detected in the event stream , adjusting the time step to a second time step, the second time step being smaller than the initial first time step.

In some possible implementation manners, the event detection result includes whether event information is detected in the event stream, wherein the step adjustment module is configured to: no event information is detected in the event stream, and When the current time step is the second time step, the time step is adjusted to the initial first time step, and the second time step is smaller than the first time step.

In some possible implementation manners, the event stream includes multiple areas, and the event detection result includes whether an event is detected in each area of the event stream, wherein the step adjustment module is configured to: When there is a target area in the plurality of areas, adjust the time step of the target area to a second time step, the target area is an area where event information is detected, and the second time step is less than The initial first time step.

In some possible implementation manners, the event detection result includes whether event information is detected in the event flow, and the event density of the event flow, wherein the step adjustment module includes: an adjustment submodule configured to in case event information is detected in the event stream, adjusting the time step according to the event density,

Wherein, the adjustment submodule is configured to: adjust the time step to a second time step when the event density is greater than or equal to a first density threshold and less than or equal to a second density threshold, The second density threshold is greater than the first density threshold, and the second time step is smaller than the initial first time step; or if the event density is greater than the second density threshold, the time step is The length is adjusted to a third time step, and the third time step is smaller than the second time step.

In some possible implementations, the event detection module is configured to: perform event detection on the event flow, and determine the current event density of the event flow; when the event density is greater than or equal to a third density threshold Next, it is determined that the event detection result is: event information is detected in the event flow.

In some possible implementations, the event detection module is configured to: divide the event stream into multiple event frames according to a preset duration; Detect to obtain the event detection result.

In some possible implementation manners, the event flow processing module is configured to: use the time step to perform brain simulation processing on the event flow through the spiking neural network, and obtain the The response information of the processing result includes the response information, wherein the differential expressions of the neuron parameters in the brain simulation are adjusted according to the time step.

In some possible implementation manners, the event stream is an event stream of a target scene collected by an event collection device, and the event stream is used to represent brightness change information of the target scene.

Fig. 4 is a block diagram of an electronic device provided by an embodiment of the present disclosure.

Referring to FIG. 4 , an embodiment of the present disclosure provides an electronic device, which includes: at least one processor 501; and a memory 502 communicatively connected to at least one processor 501; wherein, the memory 502 stores information that can be processed by at least one processor. One or more computer programs executed by the processor 501, and the one or more computer programs are executed by at least one processor 501, so that at least one processor 501 can execute the above event processing method.

In addition, an embodiment of the present disclosure also provides a computer-readable medium on which a computer program is stored, wherein the computer program implements the above event processing method when executed by a processor.

Those of ordinary skill in the art can understand that all or some of the steps in the methods disclosed above, the functional modules/units in the system, and the device can be implemented as software, firmware, hardware, and an appropriate combination thereof. In a hardware implementation, the division between functional modules/units mentioned in the above description does not necessarily correspond to the division of physical components; for example, one physical component may have multiple functions, or one function or step may be composed of several physical components. Components cooperate to execute. Some or all of the physical components may be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit . Such software may be distributed on computer readable media, which may include computer storage media (or non-transitory media) and communication media (or transitory media). As known to those of ordinary skill in the art, the term computer storage media includes both volatile and nonvolatile media implemented in any method or technology for storage of information, such as computer readable instructions, data structures, program modules, or other data. permanent, removable and non-removable media. Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disk (DVD) or other optical disk storage, magnetic cartridges, tape, magnetic disk storage or other magnetic storage devices, or can Any other medium used to store desired information and which can be accessed by a computer. In addition, as is well known to those of ordinary skill in the art, communication media typically embodies computer readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media .

Example embodiments have been disclosed herein, and while specific terms have been employed, they are used and should be construed in a generic descriptive sense only and not for purposes of limitation. In some instances, it will be apparent to those skilled in the art that features, characteristics and/or elements described in connection with a particular embodiment may be used alone, or may be described in combination with other embodiments, unless explicitly stated otherwise. Combinations of features and/or elements. Accordingly, it will be understood by those of ordinary skill in the art that various changes in form and details may be made without departing from the scope of the present disclosure as set forth in the appended claims.

Claims (12)

1. An event processing method, comprising:

acquiring an event stream to be processed;

performing event detection on the event stream, and determining the current event detection result of the event stream;

according to the event detection result, adjusting the time step for processing the event stream;

and processing the event stream by adopting the time step to obtain a processing result of the event stream.

2. The method of claim 1, wherein the event detection result includes whether event information is detected in the event stream, wherein the adjusting a time step for processing the event stream according to the event detection result includes:

And in the event that event information is detected in the event stream, adjusting the time step to be a second time step, wherein the second time step is smaller than the initial first time step.

3. The method according to claim 1 or 2, wherein the event detection result comprises whether event information is detected in the event stream, wherein the adjusting the time step of processing the event stream according to the event detection result comprises:

and under the condition that no event information is detected in the event stream and the current time step is a second time step, adjusting the time step to be an initial first time step, wherein the second time step is smaller than the first time step.

4. The method of claim 1, wherein the event stream comprises a plurality of regions, the event detection result comprises whether event information is detected in each region of the event stream,

wherein, according to the event detection result, adjusting a time step for processing the event stream includes:

and when the target area exists in the plurality of areas, adjusting the time step of the target area to be a second time step, wherein the target area is the area where the event information is detected, and the second time step is smaller than the initial first time step.

5. The method of claim 1, wherein the event detection result comprises whether event information is detected in the event stream, and an event density of the event stream,

wherein, according to the event detection result, adjusting a time step for processing the event stream includes: in case event information is detected in the event stream, adjusting the time step according to the event density,

wherein adjusting the time step according to the event density comprises:

when the event density is greater than or equal to a first density threshold and less than or equal to a second density threshold, adjusting the time step to a second time step, wherein the second density threshold is greater than the first density threshold, and the second time step is less than the initial first time step; or (b)

And when the event density is greater than a second density threshold, adjusting the time step to a third time step, wherein the third time step is smaller than the second time step.

6. The method according to any one of claims 1-5, wherein the performing event detection on the event stream to determine a current event detection result of the event stream includes:

Performing event detection on the event stream, and determining the current event density of the event stream;

and under the condition that the event density is greater than or equal to a third density threshold value, determining that the event detection result is: event information is detected in the event stream.

7. The method according to any one of claims 1-5, wherein the performing event detection on the event stream to determine a current event detection result of the event stream includes:

dividing the event stream into a plurality of event frames according to preset duration;

and carrying out event detection on the current event frame of the event stream through an event detection network to obtain the event detection result.

8. The method according to any one of claims 1-7, wherein processing the event stream with the time step to obtain a processing result of the event stream comprises:

performing brain simulation processing on the event stream by using the time step through a pulse neural network to obtain response information of the pulse neural network for the event stream, wherein the processing result comprises the response information,

and the differential expression of the neuron parameters in the brain simulation is adjusted according to the time step.

9. The method according to any of claims 1-8, wherein the event stream is an event stream of a target scene acquired by an event acquisition device, the event stream being used to characterize brightness change information of the target scene.

10. An event processing apparatus, comprising:

the event stream acquisition module is used for acquiring an event stream to be processed;

the event detection module is used for carrying out event detection on the event stream and determining the current event detection result of the event stream;

the step length adjusting module is used for adjusting the time step length for processing the event stream according to the event detection result;

and the event stream processing module is used for processing the event stream by adopting the time step to obtain the processing result of the event stream.

11. An electronic device, comprising:

at least one processor; and

a memory communicatively coupled to the at least one processor; wherein,,

the memory stores one or more computer programs executable by the at least one processor to enable the at least one processor to perform the event processing method of any of claims 1-9.

12. A computer readable medium, on which a computer program is stored, which, when being executed by a processor, implements the event processing method according to any of claims 1-9.

Images