Automatic flicker mitigation for digital display
The described system automatically mitigates digital display flicker using a flicker detection subsystem and reinforcement learning to update calibration policies, addressing the inefficiencies of traditional methods and improving display performance and lifespan.
Patent Information
- Application Number
- PCT/CN2025/092757
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2025-05-05
- Publication Date
- 2026-11-12
AI Technical Summary
Digital displays suffer from flicker conditions that traditional calibration methods fail to address effectively, particularly those that emerge over time due to aging components and environmental factors, requiring tedious manual adjustments.
A computing device employs a flicker detection subsystem using rule-based filtering and deep learning to analyze image frames, coupled with a reinforcement learning algorithm to iteratively update the display's calibration policy, automatically mitigating flicker by adjusting multiple calibration parameters.
The system provides real-time, efficient, and accurate flicker mitigation, enhancing user experience and display longevity by adapting to changing conditions without user intervention.
Smart Images

Figure CN2025092757_12112026_PF_FP_ABST
Abstract
Description
AUTOMATIC FLICKER MITIGATION FOR DIGITAL DISPLAYBACKGROUND
[0001] Digital displays can be used to provide visual representations of computer data. For instance, digital displays may include hardware elements using transmissive or emissive technologies to emit light in controlled, spatially modulated patterns to thereby display imagery.SUMMARY
[0002] This Summary is provided to introduce a selection of concepts in a simplified form that are further described below in the Detailed Description. This Summary is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to be used to limit the scope of the claimed subject matter. Furthermore, the claimed subject matter is not limited to implementations that solve any or all disadvantages noted in any part of this disclosure.
[0003] A method for flicker mitigation includes, at a computing device, receiving a plurality of image frames depicting a digital display, wherein the digital display is affected by an ongoing display flicker condition. The plurality of image frames are input to a flicker detection subsystem configured to output a severity indication that indicates an estimated severity of the display flicker condition. The severity indication is input to a calibration adjustment subsystem configured to iteratively update an existing calibration policy defined by one or more calibration parameters of the digital display, wherein the calibration adjustment system implements a reinforcement learning algorithm that is rewarded for calibration parameter updates that reduce the severity indication. Based at least in part on determining that the calibration parameter updates have reduced the severity indication to below a threshold, an updated calibration policy is output based at least in part on the calibration parameter updates.BRIEF DESCRIPTION OF THE DRAWINGS
[0004] FIG. 1 schematically shows an example computing device configured to perform automatic display flicker mitigation.
[0005] FIG. 2 schematically illustrates an example arrangement in which a camera system is used to capture image frames depicting a flicker condition.
[0006] FIG. 3 schematically illustrates another example arrangement in which a camera system is used to capture image frames depicting a flicker condition.
[0007] FIG. 4 schematically illustrates data used in training a flicker detection subsystem.
[0008] FIG. 5 schematically illustrates the use of reinforcement learning to iteratively update a calibration policy.
[0009] FIG. 6 illustrates an example method for display flicker mitigation.
[0010] FIG. 7 schematically shows an example computing system.DETAILED DESCRIPTION
[0011] Digital displays can sometimes be affected by undesirable flicker conditions. Display flicker may be caused or exacerbated by various factors, such as unstable backlight modulation, changes in the display's refresh rate, aging of display components, fluctuations in power supply, environmental conditions, etc. Display flicker refers to a phenomenon in which the brightness of a display panel fluctuates in an undesirable manner. In some cases, flicker may manifest or become increasingly pronounced over time as display components age. Environmental factors such as ambient temperature, humidity, and / or power supply stability may additionally or alternatively contribute to the emergence or amplification of display flicker. Traditional approaches to flicker mitigation often rely on factory calibration procedures that establish baseline settings during manufacturing. However, this may fail to address display flicker conditions that emerge over time. In some cases, manual calibration routines are provided to end users, allowing for post-deployment adjustment of brightness, contrast, and / or refresh rate parameters. However, these may be tedious and time consuming for a user to perform.
[0012] Accordingly, the present disclosure is directed to techniques for automatic display flicker mitigation. First, a computing device detects that a digital display is affected by an ongoing display flicker condition based at least in part on a plurality of image frames depicting the digital display. The image frames may be captured by a camera system that is integrated within, or external to, the computing device, as will be described in more detail below. The image frames are input to a flicker detection subsystem of the computing device, which is configured to output a severity indication that indicates the estimated severity of the display flicker condition. The flicker detection subsystem may be implemented in any suitable way, using any suitable combination of computer hardware, firmware, and / or software. In some examples, the flicker detection subsystem is implemented via a suitable processor configured to instantiate and perform one or more suitable algorithms and / or machine learning models. The flicker detection system may use a suitable combination of rules-based filtering and / or deep learning techniques to generate the severity indication, as will be described below. Next, a calibration adjustment system is used to iteratively update an existing calibration policy of the digital display. The calibration adjustment system implements a reinforcement learning algorithm that is rewarded for policy updates that reduce the severity indication. When the severity indication has been reduced to below a threshold, indicating that the flicker condition has been mitigated or alleviated, the computing device outputs an updated calibration policy for the digital display. For instance, the updated calibration policy may be stored in firmware of the digital display, and used to control the display's behavior going forward.
[0013] In this manner, the techniques described herein can beneficially be used to mitigate display flicker conditions automatically, in real-time. This improves human-computer interaction and reduces the burden of user input to the computing device. As a result, the displayed image content is presented more faithfully and without undesirable flicker effects, without requiring the user to perform tedious manual calibration steps. Use of deep-learning and / or reinforcement learning may enable more accurate and efficient display calibration, allowing multiple calibration parameters to be adjusted at once in response to diverse display flickering effects. The techniques described herein may beneficially extend the useable lifespan of the digital display, by automatically reducing perceptible display flicker over time as components of the display age, and / or as environmental conditions change. The techniques described herein may beneficially be applied to a range of different suitable display technologies, thereby supporting cross-platform applicability. Furthermore, these techniques may facilitate the use of lower-power or energy saving modes with the display device, which may otherwise contribute to increased display flickering, thereby reducing the overall power expenditure by the display device.
[0014] FIG. 1 schematically shows an example computing device 100 that may be used to implement any or all of the techniques described herein. A “computing device” according to the present disclosure may have any suitable capabilities, hardware configuration, and form factor. In some examples, computing device 100, and / or any of the other computing devices described herein, may be implemented as computing system 700 described below with respect to FIG. 7.
[0015] In FIG. 1, computing device 100 is communicatively coupled with a digital display 102. As used herein, a “digital display” refers to any suitable collection of hardware components useable to render visual content by selectively activating discrete image-forming elements (e.g., pixels) based at least in part on control signals received from computing device 100. In this manner, the display device may be used to present graphical user interfaces (GUIs) , image content, video content, multimedia presentations, and / or any other suitable visual content. A digital display may include, for example, a liquid crystal display (LCD) , organic light-emitting diode (OLED) display, light-emitting diode (LED) display, microLED display, electronic ink (e-ink) display, and / or other suitable display technology. The visual output of a digital display may be generated in response to a stream of digital image data that defines the color, brightness, and / or other characteristics of individual pixels of the display. Digital displays may incorporate additional subsystems such as backlighting units, timing controllers, driver circuitry, etc.
[0016] In FIG. 1, the behavior of the digital display is at least partially influenced by a calibration policy 104, which itself is defined by a plurality of calibration parameters 106. As used herein, a "calibration policy" refers to a collection of parameters that collectively define the operational configuration of a display device. The calibration policy governs how the display renders image content under existing conditions. Any modification to one or more calibration parameters constitutes an update to the calibration policy. In this manner, updates to the calibration policy can beneficially be used to adapt the behavior of the display device to attempt to mitigate a detected flicker condition, which may arise due to aging components, environmental conditions, usage patterns, and / or other changes in system performance. In this manner, device performance and user comfort can be enhanced through automated updates to the calibration policy.
[0017] Similarly, a "calibration parameter" refers to an individual configurable variable, threshold, coefficient, or other suitable value within the calibration policy that affects the display′s output and / or behavior. Calibration parameters may be updated periodically or continuously as part of an automatic or semi-automatic calibration process. Non-limiting examples of calibration parameters include a pulse-width modulation (PWM) frequency of the display, a duty cycle of the display, a refresh rate of the display, a phase compensation of the display, etc. The techniques described herein may beneficially enable two or more of these parameters to be adjusted at once, improving the speed and efficiency of the calibration process.
[0018] In the example of FIG. 1, computing device 100 is additionally communicatively coupled with a camera system 108. The camera system is used to capture a plurality of image frames 110 depicting display device 102. In some cases, the computing device may additionally or alternatively receive other types of data, such as backlight signal waveforms and / or pulse width modulation (PWM) signals. Such data may be derived from the plurality of image frames, and / or received separately. The camera system may have any suitable relationship with respect to the computing device and display device. In various examples, any or all of these components may be integrated into the same physical housing, and / or may be retained within their own separate respective housings. Example arrangements will be described below with respect to FIGS. 2 and 3.
[0019] FIG. 2 schematically illustrates an example scenario in which a camera system is used to capture image frames depicting a digital display. Specifically, FIG. 2 schematically shows an example computing device 200, which includes a digital display 202. The digital display is affected by an ongoing display flicker condition. A camera system 204 is used to capture a plurality of image frames 206 depicting the display device 202. These image frames may beneficially be used by the computing device to detect the ongoing flicker condition, and may cause the computing device to initiate a flicker mitigation process to automatically resolve the display flicker condition, as will be described in more detail below.
[0020] In this example, the digital display is integrated into the same physical housing as the computing device, while the camera system is an external camera system communicatively coupled with the computing device, but housed in a separate physical housing from the computing device. It will be understood that this arrangement is non-limiting. For instance, in other implementations, the display device and computing device may be disposed in physically separate housings -e.g., the computing device may take the form of a desktop computer, which is used in tandem with a separate computer monitor. In cases where the digital display and / or the camera system are in separate housings from the computing device, they may be communicatively coupled with the computing device using any suitable wired and / or wireless communication interfaces. In this manner, the techniques described herein beneficially may be adapted to a wide range of different hardware configurations.
[0021] The plurality of image frames may include any suitable quantity of image frames, which may be captured at any suitable time and with any suitable frequency. The image frames may have any suitable characteristics, such as resolution, encoding, color space, etc. In some examples, the image frames may be individual frames of a digital video captured by the camera system. In other examples, the image frames may comprise a sequence of still images acquired at irregular intervals and / or in response to specific triggering conditions, such as changes in luminance, refresh rate transitions, suspected flicker events, etc. The image frames may be analyzed individually or in aggregate to detect indicators of display flicker. The frequency, resolution, and capture mode of the image frames may be selected based on system constraints, processing resources, the desired sensitivity of flicker detection, and / or other suitable factors.
[0022] Similarly, the camera system itself may have any suitable capabilities and characteristics. For instance, in some cases, the camera system is integrated into a separate computing device, such as a smartphone or tablet operated by a user. In such configurations, the camera system may be manually aimed at the display to capture image data for diagnostic and / or calibration purposes. Such scenarios can beneficially reduce the cost of the system, by enabling an existing camera (e.g., in the user's smartphone) to be repurposed for display calibration. In other cases, the camera system may take another suitable form, such as a standalone camera having a fixed position (e.g., such as a surveillance camera mounted to a wall or ceiling) , which can beneficially simplify display calibration from the user's perspective. The camera system may support various frame rates, exposure settings, and / or spectral sensitivities, which may be selectively adjusted to enhance the detection of flicker or other temporal anomalies. In some embodiments, the camera system may further include image processing circuitry or software configured to perform preliminary analysis, filtering, and / or compression of captured data prior to transmission of the image frames to the computing device, which can beneficially conserve processing resources of the computing device for other tasks.
[0023] FIG. 3 schematically illustrates another example scenario in which a camera system is used to capture image frames depicting a digital display. Specifically, FIG. 3 schematically shows another example computing device 300, including an integrated digital display 302. In this example, the computing device additionally includes an integrated camera system 304, which is used to capture image frames depicting the digital display. In other words, in this example, the plurality of image frames are captured by a camera system that is also integrated into the same physical housing as the computing device. This is achieved through the use of a reflective surface 306 (e.g., a suitable mirror) , which facilitates visibility of the digital display in the images captured by camera system 304. The arrangement between the computing device and reflective surface may be manually established by a user (e.g., by moving the display device and / or reflective surface until a suitable spacing and orientation is achieved) . In some cases, the display device and reflective surface may have a fixed relationship -e.g., in cases where the display device is implemented in a terminal that is rarely or never moved.
[0024] In some examples, the integrated computing system may be capable of imaging the display device without the need for an external reflective surface. For instance, the camera system and / or display device may include suitable internal optical structures enabling the camera device to capture images of the display device. These may include, for instance, partially reflective coatings, beam-splitting elements, waveguides, and / or prisms configured to redirect a portion of the display's emitted light toward the camera sensor. In some implementations, a light path may be formed within the device enclosure using angled surfaces, optical tunnels, and / or other suitable optical elements that allow the camera system to receive a representative view of the display content. In some examples, the computing device may directly monitor backlight control signals, in addition to or instead of receiving image frames depicting the digital display. Such configurations may beneficially facilitate continuous or periodic monitoring of display behavior in a compact form factor, without requiring external hardware or manual alignment.
[0025] Regardless of the specific arrangement, the computing device receives a plurality of image frames depicting the digital display. These image frames can then be processed to detect and automatically mitigate an ongoing display flicker condition. This will be described with reference to FIG. 1. As discussed above, in FIG. 1, computing device 100 is communicatively coupled with digital display 102 and camera system 108. Computing device 100 receives a plurality of image frames 110 depicting digital display 102, which is affected by an ongoing display flicker system.
[0026] The plurality of image frames are then input to a flicker detection subsystem 112, which is configured to output a severity indication that indicates an estimated severity of the display flicker condition. The flicker detection subsystem may use any suitable combination of rules-based filtering and / or deep-learning techniques to generate the severity indication. In the example of FIG. 1, the flicker detection subsystem includes a rule-based flickering step 114, a spatial feature extraction step 116, and a temporal dynamics step 118. In other examples, one or more of these steps may be omitted, and / or additional steps may be included. Use of deep-learning technologies (e.g., such as in the spatial feature extraction and / or temporal dynamics steps) can improve the ability of the system to adapt to changing display conditions over time -e.g., as components age. This improves the user experience by improving the accuracy with which imagery is displayed, and improves the longevity of the display device.
[0027] In FIG. 1, the rule-based flickering step includes a time-domain filter 120 and a frequency-domain filter 122. These may be implemented in any suitable way in order to detect and isolate indicators of potential flickering in the input image frames. For instance, the time-domain filter 120 may analyze temporal changes in pixel intensity across successive frames to identify periodic variations exceeding a defined threshold, which may indicate brightness instability. The frequency-domain filter 122 may apply a transform, such as a fast Fourier transform (FFT) , to derive the spectral content of intensity fluctuations over time, thereby isolating frequency components consistent with known flicker patterns (e.g., modulation at or near the backlight PWM frequency) . In some embodiments, the outputs of both filters may be combined or weighted to produce a composite flicker likelihood score, which may then be compared against one or more thresholds to trigger a calibration policy update and / or initiate other suitable mitigation actions. This rule-based approach may beneficially enable rapid, low-latency identification of potential flicker conditions without requiring extensive computational resources, making it well-suited for real-time or resource-constrained display environments. In particular, the use of both time-domain and frequency-domain filtering can detect and mitigate a wide variety of different potential flickering conditions having various potential causes, improving the efficacy of the flicker mitigation techniques and improving overall display performance.
[0028] In FIG. 1, the spatial feature extraction step 116 implements a first neural network 124. The first neural network may take any suitable form, and may be trained to perform spatial feature extraction in any suitable way. As one non-limiting example, the first neural network may be implemented as a convolutional neural network (CNN) . The CNN may be configured to receive one or more image frames as input and generate output data that highlights spatial patterns, regions of interest, and / or structural anomalies in the image frames that may correlate with flicker-related artifacts. Training of the CNN may involve supervised or unsupervised learning using labeled datasets that include examples of known flicker conditions, beneficially allowing the network to learn spatial cues indicative of display instability and improving the ability of the computing device to detect and respond to display flicker.
[0029] The temporal dynamics step implements a second neural network 126. As with the first neural network, the second neural network may take any suitable form and may be trained to analyze temporal dynamics in input image frames in any suitable way. As one non-limiting example, the second neural network may be implemented as a long short-term memory (LSTM) network. This may be advantageous, as it is believed that LSTM architectures are well suited for analysis of time-series data. As another example, a recurrent neural network (RNN) may be used. In either case, the second neural network may be configured to receive sequences of pixel values and / or other temporally ordered data derived from the image frames, and to identify patterns indicative of periodic visual fluctuations consistent with display flicker. The network beneficially may be trained to distinguish between benign temporal changes, such as those caused by dynamic content transitions, and anomalous temporal signatures associated with display flickering, thereby improving the ability of the system to detect genuine display flicker as opposed to changes in brightness that are normal and intended.
[0030] More generally, the flicker detection subsystem may be trained in any suitable way. In some examples, the flicker detection subsystem is previously trained using a plurality of training image frames, which may be representative of various different types of flickering effects that a display may exhibit, in addition to image frames depicting normal display behavior. These training image frames may be accompanied with corresponding training labels indicating whether the training image frames exhibit display flicker. The training image frames may additionally include any other suitable data, such as indications of the type of flicker that is exhibited, and / or parameters defining the nature of the flicker (e.g., frequency, cycle length, spatial distribution, etc. ) , as this may beneficially enhance the ability of the trained flicker detection subsystem to detect and respond to various types of display flicker.
[0031] This is schematically illustrated with respect to FIG. 4. Specifically, FIG. 4 schematically depicts training of a flicker detection subsystem 400. As shown, the flicker detection subsystem is trained based at least in part on a training dataset 402, which includes a plurality of training image frames, along with corresponding training labels 406 indicating whether the training image frames exhibit display flicker. The flicker detection subsystem processes such data and generates a set of output data 408. Over a plurality of training iterations, parameters of the one or more neural networks of the flicker detection subsystem may be iteratively updated until the training output is consistent with the training labels in the training dataset. In this manner, the flicker detection subsystem may be trained to detect flicker in runtime images of a digital display, as discussed above. In some cases, some to all of the training process may occur offiine, while inference and / or fine tuning may occur online.
[0032] Returning again to FIG. 1, the flicker detection subsystem generates output data 128 based at least in part on the plurality of input image frames 110. As shown, this includes a severity indication 130 that indicates an estimated severity of the display flicker condition. In this example, the flicker detection subsystem additionally outputs a plurality of flicker parameters 132 defining characteristics of the display flicker condition. These may include any of a wide variety of different types of data that can be used to quantify different aspects of the observed flicker. As non-limiting examples, the flicker parameters may include at least one of a frequency of the display flicker condition, an amplitude of the display flicker condition, a periodicity of the display flicker condition, and localized brightness variations detected on the digital display. By characterizing the display flicker according to various different types of parameters, the applied flicker mitigation operations can be more carefully tailored to the specific type of flicker that is observed, which beneficially improves the ability of the computing system to mitigate the observed flicker effectively.
[0033] The severity indication 130 and the flicker parameters 132 may be generated by the flicker detection subsystem based on any suitable combination or weighting of the outputs of the rule-based filtering step, the spatial feature extraction step, and / or the temporal dynamics step. In some embodiments, the flicker detection subsystem may apply a fusion algorithm or decision logic to aggregate information from these different analysis stages, thereby producing a unified assessment of flicker presence and severity. For instance, higher confidence may be assigned to temporal patterns identified by the second neural network when corroborated by frequency-domain signals from the rule-based filtering step, or localized spatial anomalies detected by the first neural network may be used to refine the severity indication when time-domain fluctuations are ambiguous. The weighting and contribution of each component may be fixed or adaptive, and may vary depending on display type, operational conditions, and / or historical flicker data. This may beneficially enable more robust and context-sensitive detection, particularly in environments where one analysis domain may be insufficient to independently characterize the flicker behavior.
[0034] Continuing with FIG. 1, the severity indication is input to a calibration adjustment subsystem 134. As with the flicker detection subsystem, the calibration adjustment subsystem may be implemented in any suitable way, using any suitable combination of computer hardware, firmware, and / or software. For instance, in some examples, the calibration adjustment subsystem may be implemented as a computer processor that executes one or more suitable algorithms and / or other suitable processes, such as a reinforcement learning algorithm. The calibration adjustment subsystem is configured to iteratively update the existing calibration policy 104 of the digital display by updating one or more of the calibration parameters 106 defining the calibration policy. The calibration adjustment system implements a reinforcement learning algorithm 13 6 that is rewarded for calibration parameter updates that reduce the severity indication. In other words, upon detecting display flicker, the calibration adjustment system may apply one or more changes to the calibration policy, to attempt to improve or alleviate the flicker condition. As the computing device continues to receive successive image frames, and analyze such image frames using the flicker detection subsystem, the changes applied by the calibration adjustment subsystem may result in increases or decreases to the severity indication. Ifthe severity indication decreases, the reinforcement learning algorithm may receive a positive reward, incentivizing the algorithm to perform similar changes to those that improved the flicker condition. Ifthe severity indication increases, the reinforcement learning algorithm may in some cases receive a negative reward, disincentivizing similar changes. In this manner, use of a reward provides feedback to the reinforcement learning algorithm, which beneficially improves the speed at which the algorithm can discover and implement parameter updates that improve or resolve the display flicker condition.
[0035] The reinforcement learning algorithm may be implemented in any suitable way and take any suitable form. In some examples, the reinforcement learning algorithm may be implemented as one of a deep-Q network (DQN) or a proximal policy optimization (PPO) algorithm. These may beneficially enable the calibration adjustment system to learn effective parameter update strategies through repeated interaction with the display environment, using observed outcomes, such as changes in flicker severity or system performance, as feedback. In this manner, the reinforcement learning algorithm may beneficially serve to iteratively improve the display flicker condition based on the data output by the flicker detection subsystem, such as the severity indication and / or the flicker parameters.
[0036] Additionally, or alternatively, the computing device may receive user feedback indicating whether the flicker condition has improved. In the example of FIG. 1, the computing device additionally receives user feedback 138, which is provided to the calibration adjustment subsystem. This feedback may take any suitable form and may be provided in any suitable way -e.g., such as via a software application executed by the computing device. The user feedback may indicate whether the flicker condition has improved, worsened, stayed the same, provide a numerical estimate of the current severity of the flicker (e.g., on a scale of one to ten) , etc. In some cases, the reinforcement learning algorithm may be rewarded based at least in part on the user input indicating that the display flicker condition has improved. This can beneficially enable flicker reduction based on a user's perception of the flicker, which may in some cases differ from the computing device's characterization of the flicker derived from the plurality of image frames.
[0037] The specific calibration parameter updates applied by the reinforcement learning algorithm may be determined in any suitable way. In some cases, the calibration parameter updates applied by the calibration adjustment system may be determined based at least in part on the plurality of flicker parameters generated by the flicker detection subsystem. For example, a detected flicker frequency may inform modifications to the display′s backlight modulation rate or refresh cycle timing. The reinforcement learning algorithm may evaluate the effectiveness of past calibration actions based on observed changes in flicker parameters over time and iteratively refine its policy to reduce future flicker severity. In some implementations, the algorithm may explore new calibration parameter values during operation, balancing the potential improvement in flicker conditions against the risk of introducing new visual artifacts or system instability. Updates to the calibration policy may be applied continuously or in discrete steps, and may be constrained by predefined limits or profiles associated with particular display types, usage modes, and / or environmental conditions. This may beneficially enable the reinforcement learning algorithm to adapt to the specific types of display flicker currently being observed, which can improve the efficacy of the calibration adjustment system at reducing or eliminating display flicker over time.
[0038] This process of iterative calibration policy updates may continue for any suitable length of time and / or any suitable number of iteration cycles. In some examples, the calibration adjustment subsystem may continue performing policy updates until the severity indication is below a predefined threshold. In other words, based at least in part on determining that the calibration parameter updates have reduced the severity indication to below a threshold, the calibration adjustment subsystem may output an updated calibration policy based at least in part on the calibration parameter updates. As used herein, the updated policy may be “output” in any suitable way, such as by saving the updated policy to a configuration file, writing it to computer storage, transmitting it over a computer network, etc. In some examples, outputting the updated calibration policy includes replacing the existing calibration policy in firmware of the display device, such that the updated calibration policy is used to control the behavior of the digital display going forward. In this manner, the existing calibration policy may beneficially be replaced by an updated calibration policy that reduces or alleviates the observed flickering.
[0039] FIG. 5 schematically illustrates the use of a reinforcement learning algorithm in more detail. Specifically, as shown, FIG. 5 schematically depicts a reinforcement learning algorithm 500, which generates policy updates 502 that are applied to a calibration policy 504 of a digital display. Changes to the calibration policy affect a display context 506. This is influenced by the current set of calibration parameters 508 (such as the display's PWM frequency, duty cycle, refresh rate, phase compensation, etc. ) , a severity indication 510 output by a flicker detection subsystem, flicker parameters 512 output by the flicker detection subsystem, and / or human feedback 514 provided by a human user. Consideration of a diverse set of calibration parameters can provide both more insight and more control over the current performance state of the digital display, improving the ability of the system to detect and effectively respond to display flicker. Based at least in part on such data, a reward 516 is provided to the reinforcement learning algorithm, which in turn affects subsequent policy updates applied by the reinforcement learning algorithm. For instance, policy updates that reduce the severity indication may result in a positive reward, while policy updates that increase the severity indication may result in a negative reward. This process may continue for any suitable length of time -e.g., until the severity indication falls below a threshold, until the computing device receives human feedback indicating that the flickering has ceased, and / or another process terminating condition is met.
[0040] FIG. 6 illustrates an example method 600 for automatic display flicker mitigation. Steps of method 600 may be initiated, terminated, and / or repeated at any suitable time and in response to any suitable condition. One or more steps of method 600 may be performed by a suitable computing system of one or more computing devices. A computing device implementing steps of method 600 may have any suitable capabilities, hardware configuration, and form factor. In some examples, method 600 may be performed by computing system 700 described below with respect to FIG. 7.
[0041] At 602, method 600 includes receiving a plurality of image frames depicting a digital display affected by a display flicker condition. The image frames may be received from a camera system configured to capture visual output from the display under normal operating conditions. The image frames may be acquired in real time or retrieved from a previously stored data source, and may include frames captured at a fixed or variable frame rate. The received frames may include visual evidence of flicker artifacts, which may be analyzed in subsequent steps to assess the severity and characteristics of the flicker condition.
[0042] At 604, method 600 includes inputting the plurality of image frames to a flicker detection subsystem configured to output a severity indication. The flicker detection subsystem may comprise one or more processing components, including rule-based filters, neural network models, and / or other analytical modules, as described above. The subsystem analyzes the temporal and spatial characteristics of the image frames to identify indicators of flicker, and produces a severity indication that reflects the estimated intensity, persistence, and / or visual impact of the detected flicker condition. In some embodiments, the severity indication may be represented as a numerical score, a categorical level (e.g., low, moderate, high) , and / or a normalized value bounded within a defined range. The severity indication may be used alone or in conjunction with other outputs, such as extracted flicker parameters, to inform downstream calibration or mitigation operations.
[0043] At 606, method 600 includes inputting the severity indication to a calibration adjustment subsystem configured to iteratively update an existing calibration policy. The calibration adjustment subsystem may evaluate the severity indication in the context of current calibration parameters, recent update history, and / or other system state information to determine whether and how the calibration policy should be modified. In some implementations, the subsystem may utilize a reinforcement learning algorithm, such as a deep-Q network or PPO algorithm, to select calibration parameter updates that are expected to reduce flicker severity while maintaining overall display performance. The calibration adjustment subsystem may apply updates incrementally, allowing for continuous adaptation to evolving flicker conditions without introducing abrupt changes in display output. In this manner, the calibration policy may beneficially be refined over time to improve stability and mitigate flicker effects under a variety of operational and environmental conditions.
[0044] At 608, method 600 includes determining whether the severity indication has been reduced to below a predefined threshold. Ifno, method 600 returns to 606, where the calibration adjustment subsystem continues iteratively updating the calibration policy based on newly observed severity indications and / or other feedback. This loop may continue until the severity indication stabilizes within an acceptable range, indicating that the display flicker condition has been sufficiently mitigated. If yes, method 600 proceeds to 610, which includes outputting an updated calibration policy. The updated calibration policy reflects the most recent set of calibration parameters determined to effectively reduce flicker under current conditions, and may be stored, transmitted, and / or applied directly to the display device for real-time implementation. In some embodiments, the updated calibration policy may also be logged or used to inform future calibration strategies under similar operating scenarios.
[0045] The methods and processes described herein may be tied to a computing system of one or more computing devices. In particular, such methods and processes may be implemented as an executable computer-application program, a network-accessible computing service, an application-programming interface (API) , a library, or a combination of the above and / or other compute resources.
[0046] FIG. 7 schematically shows a simplified representation of a computing system 700 configured to provide any to all of the compute functionality described herein. Computing system 700 may take the form of one or more personal computers, network-accessible server computers, tablet computers, home-entertainment computers, gaming devices, mobile computing devices, mobile communication devices (e.g., smart phone) , virtual / augmented / mixed reality computing devices, wearable computing devices, Internet of Things (IoT) devices, embedded computing devices, and / or other computing devices.
[0047] Computing system 700 includes a logic subsystem 702 and a storage subsystem 704. Computing system 700 may optionally include a display subsystem 706, input subsystem 708, communication subsystem 710, and / or other subsystems not shown in FIG. 7.
[0048] Logic subsystem 702 includes one or more physical devices configured to execute instructions. For example, the logic subsystem may be configured to execute instructions that are part of one or more applications, services, or other logical constructs. In some examples, either or both of the flicker detection subsystem and calibration adjustment subsystem may be implemented by the logic subsystem. The logic subsystem may include one or more hardware processors configured to execute software instructions. Additionally, or alternatively, the logic subsystem may include one or more hardware or firmware devices configured to execute hardware or firmware instructions. Processors of the logic subsystem may be single-core or multi-core, and the instructions executed thereon may be configured for sequential, parallel, and / or distributed processing. Individual components of the logic subsystem optionally may be distributed among two or more separate devices, which may be remotely located and / or configured for coordinated processing. Aspects of the logic subsystem may be virtualized and executed by remotely-accessible, networked computing devices configured in a cloud-computing configuration.
[0049] Storage subsystem 704 includes one or more physical devices configured to temporarily and / or permanently hold computer information such as data and instructions executable by the logic subsystem. When the storage subsystem includes two or more devices, the devices may be collocated and / or remotely located. Storage subsystem 704 may include volatile, nonvolatile, dynamic, static, read / write, read-only, random-access, sequential-access, location-addressable, file-addressable, and / or content-addressable devices. Storage subsystem 704 may include removable and / or built-in devices. When the logic subsystem executes instructions, the state of storage subsystem 704 may be transformed -e.g., to hold different data.
[0050] Aspects of logic subsystem 702 and storage subsystem 704 may be integrated together into one or more hardware-logic components. Such hardware-logic components may include program-and application-specific integrated circuits (PASIC / ASICs) , program-and application-specific standard products (PSSP / ASSPs) , system-on-a-chip (SOC) , and complex programmable logic devices (CPLDs) , for example.
[0051] The logic subsystem and the storage subsystem may cooperate to instantiate one or more logic machines. As used herein, the term “machine” is used to collectively refer to the combination of hardware, firmware, software, instructions, and / or any other components cooperating to provide computer functionality. In other words, “machines” are never abstract ideas and always have a tangible form. A machine may be instantiated by a single computing device, or a machine may include two or more sub-components instantiated by two or more different computing devices. In some implementations a machine includes a local component (e.g., software application executed by a computer processor) cooperating with a remote component (e.g., cloud computing service provided by a network of server computers) . The software and / or other instructions that give a particular machine its functionality may optionally be saved as one or more unexecuted modules on one or more suitable storage devices.
[0052] Machines may be implemented using any suitable combination of state-of-the-art and / or future machine learning (ML) , artificial intelligence (AI) , and / or natural language processing (NLP) techniques. Non-limiting examples of techniques that may be incorporated in an implementation of one or more machines include support vector machines, multi-layer neural networks, convolutional neural networks (e.g., including spatial convolutional networks for processing images and / or videos, temporal convolutional neural networks for processing audio signals and / or natural language sentences, and / or any other suitable convolutional neural networks configured to convolve and pool features across one or more temporal and / or spatial dimensions) , recurrent neural networks (e.g., long short-term memory networks) , associative memories (e.g., lookup tables, hash tables, Bloom Filters, Neural Turing Machine and / or Neural Random Access Memory) , word embedding models (e.g., GloVe or Word2Vec) , unsupervised spatial and / or clustering methods (e.g., nearest neighbor algorithms, topological data analysis, and / or k-means clustering) , graphical models (e.g., (hidden) Markov models, Markov random fields, (hidden) conditional random fields, and / or AI knowledge bases) , and / or natural language processing techniques (e.g., tokenization, stemming, constituency and / or dependency parsing, and / or intent recognition, segmental models, and / or super-segmental models (e.g., hidden dynamic models) ) .
[0053] In some examples, the methods and processes described herein may be implemented using one or more differentiable functions, wherein a gradient of the differentiable functions may be calculated and / or estimated with regard to inputs and / or outputs of the differentiable functions (e.g., with regard to training data, and / or with regard to an objective function) . Such methods and processes may be at least partially determined by a set of trainable parameters. Accordingly, the trainable parameters for a particular method or process may be adjusted through any suitable training procedure, in order to continually improve functioning of the method or process.
[0054] Non-limiting examples of training procedures for adjusting trainable parameters include supervised training (e.g., using gradient descent or any other suitable optimization method) , zero-shot, few-shot, unsupervised learning methods (e.g., classification based on classes derived from unsupervised clustering methods) , reinforcement learning (e.g., deep Q learning based on feedback) and / or generative adversarial neural network training methods, belief propagation, RANSAC (random sample consensus) , contextual bandit methods, maximum likelihood methods, and / or expectation maximization. In some examples, a plurality of methods, processes, and / or components of systems described herein may be trained simultaneously with regard to an objective function measuring performance of collective functioning of the plurality of components (e.g., with regard to reinforcement feedback and / or with regard to labelled training data) . Simultaneously training the plurality of methods, processes, and / or components may improve such collective functioning. In some examples, one or more methods, processes, and / or components may be trained independently of other components (e.g., offline training on historical data) .
[0055] When included, display subsystem 706 may be used to present a visual representation of data held by storage subsystem 704. This visual representation may take the form of a graphical user interface (GUI) . Display subsystem 706 may include one or more display devices utilizing virtually any type of technology. In some implementations, display subsystem may include one or more virtual-, augmented-, or mixed reality displays.
[0056] When included, input subsystem 708 may comprise or interface with one or more input devices. An input device may include a sensor device or a user input device. Examples of user input devices include a keyboard, mouse, touch screen, or game controller. In some embodiments, the input subsystem may comprise or interface with selected natural user input (NUI) componentry. Such componentry may be integrated or peripheral, and the transduction and / or processing of input actions may be handled on-or off-board. Example NUI componentry may include a microphone for speech and / or voice recognition; an infrared, color, stereoscopic, and / or depth camera for machine vision and / or gesture recognition; a head tracker, eye tracker, accelerometer, and / or gyroscope for motion detection and / or intent recognition.
[0057] When included, communication subsystem 710 may be configured to communicatively couple computing system 700 with one or more other computing devices. Communication subsystem 710 may include wired and / or wireless communication devices compatible with one or more different communication protocols. The communication subsystem may be configured for communication via personal-, local-and / or wide-area networks.
[0058] The methods and processes disclosed herein may be configured to give users and / or any other humans control over any private and / or potentially sensitive data. Whenever data is stored, accessed, and / or processed, the data may be handled in accordance with privacy and / or security standards. When user data is collected, users or other stakeholders may designate how the data is to be used and / or stored. Whenever user data is collected for any purpose, the user data may only be collected with the utmost respect for user privacy (e.g., user data may be collected only when the user owning the data provides affirmative consent, and / or the user owning the data may be notified whenever the user data is collected) . If the data is to be released for access by anyone other than the user or used for any decision-making process, the user's consent may be collected before using and / or releasing the data. Users may opt-in and / or opt-out of data collection at any time. After data has been collected, users may issue a command to delete the data, and / or restrict access to the data. All potentially sensitive data optionally may be encrypted and / or, when feasible, anonymized, to further protect user privacy. Users may designate portions of data, metadata, or statistics / results of processing data for release to other parties, e.g., for further processing. Data that is private and / or confidential may be kept completely private, e.g., only decrypted temporarily for processing, or only decrypted for processing on a user device and otherwise stored in encrypted form. Users may hold and control encryption keys for the encrypted data. Alternately or additionally, users may designate a trusted third party to hold and control encryption keys for the encrypted data, e.g., so as to provide access to the data to the user according to a suitable authentication protocol.
[0059] When the methods and processes described herein incorporate ML and / or AI components, the ML and / or AI components may make decisions based at least partially on training of the components with regard to training data. Accordingly, the ML and / or AI components may be trained on diverse, representative datasets that include sufficient relevant data for diverse users and / or populations of users. In particular, training data sets may be inclusive with regard to different human individuals and groups, so that as ML and / or AI components are trained, their performance is improved with regard to the user experience of the users and / or populations of users.
[0060] ML and / or AI components may additionally be trained to make decisions so as to minimize potential bias towards human individuals and / or groups. For example, when AI systems are used to assess any qualitative and / or quantitative information about human individuals or groups, they may be trained so as to be invariant to differences between the individuals or groups that are not intended to be measured by the qualitative and / or quantitative assessment, e.g., so that any decisions are not influenced in an unintended fashion by differences among individuals and groups.
[0061] ML and / or AI components may be designed to provide context as to how they operate, so that implementers of ML and / or AI systems can be accountable for decisions / assessments made by the systems. For example, ML and / or AI systems may be configured for replicable behavior, e.g., when they make pseudo-random decisions, random seeds may be used and recorded to enable replicating the decisions later. As another example, data used for training and / or testing ML and / or AI systems may be curated and maintained to facilitate future investigation of the behavior of the ML and / or AI systems with regard to the data. Furthermore, ML and / or AI systems may be continually monitored to identify potential bias, errors, and / or unintended outcomes.
[0062] This disclosure is presented by way of example and with reference to the associated drawing figures. Components, process steps, and other elements that may be substantially the same in one or more of the figures are identified coordinately and are described with minimal repetition. It will be noted, however, that elements identified coordinately may also differ to some degree. It will be further noted that some figures may be schematic and not drawn to scale. The various drawing scales, aspect ratios, and numbers of components shown in the figures may be purposely distorted to make certain features or relationships easier to see.
[0063] In an example, a method for flicker mitigation comprises: at a computing device, receiving a plurality of image frames depicting a digital display, wherein the digital display is affected by an ongoing display flicker condition; inputting the plurality of image frames to a flicker detection subsystem configured to output a severity indication that indicates an estimated severity of the display flicker condition; inputting the severity indication to a calibration adjustment subsystem configured to iteratively update an existing calibration policy defined by one or more calibration parameters of the digital display, wherein the calibration adjustment system implements a reinforcement learning algorithm that is rewarded for calibration parameter updates that reduce the severity indication; and based at least in part on determining that the calibration parameter updates have reduced the severity indication to below a threshold, outputting an updated calibration policy based at least in part on the calibration parameter updates. In this example or any other example, the flicker detection subsystem is configured to perform a rule-based filtering step using time-domain and frequency-domain filtering, a spatial feature extraction step using a first neural network, and a temporal dynamics step using a second neural network. In this example or any other example, the first neural network is a convolutional neural network (CNN) , and the second neural network is a long short-term memory (LSTM) network. In this example or any other example, the flicker detection system additionally outputs a plurality of flicker parameters defining characteristics of the display flicker condition, and wherein the calibration parameter updates applied by the calibration adjustment system are determined based at least in part on the plurality of flicker parameters. In this example or any other example, the plurality of flicker parameters include at least one of a frequency of the display flicker condition, an amplitude of the display flicker condition, a periodicity of the display flicker condition, and localized brightness variations detected on the digital display. In this example or any other example, the method further comprises receiving user input indicating whether the display flicker condition has improved, and wherein the reinforcement learning algorithm is further rewarded based at least in part on the user input indicating that the display flicker condition has improved. In this example or any other example, the flicker detection subsystem is previously trained based at least in part on a plurality of training image frames, along with corresponding training labels indicating whether the training image frames exhibit display flicker. In this example or any other example, the digital display is integrated into a same physical housing as the computing device. In this example or any other example, the plurality of image frames are captured by a camera system that is also integrated into the same physical housing as the computing device. In this example or any other example, the plurality of image frames are captured by an external camera system communicatively coupled with the computing device, wherein the external camera system is in a separate physical housing from the computing device. In this example or any other example, outputting the updated calibration policy includes replacing the existing calibration policy in firmware of the display device. In this example or any other example, the one or more calibration parameters of the digital display include at least one of a pulse-width modulation (PWM) frequency of the digital display, a duty cycle of the digital display, a refresh rate of the digital display, and a phase compensation of the digital display. In this example or any other example, the reinforcement learning algorithm is one of a deep-Q network (DQN) or a proximal policy optimization (PPO) algorithm.
[0064] In an example, a computing device comprises: a logic subsystem; and a storage subsystem holding instructions executable by the logic subsystem to: receive a plurality of image frames depicting a digital display, wherein the digital display is affected by an ongoing display flicker condition; input the plurality of image frames to a flicker detection subsystem configured to output a severity indication that indicates an estimated severity of the display flicker condition; input the severity indication to a calibration adjustment subsystem configured to iteratively update an existing calibration policy defined by one or more calibration parameters of the digital display, wherein the calibration adjustment system implements a reinforcement learning algorithm that is rewarded for calibration parameter updates that reduce the severity indication; and based at least in part on determining that the calibration parameter updates have reduced the severity indication to below a threshold, output an updated calibration policy based at least in part on the calibration parameter updates. In this example or any other example, the flicker detection subsystem is configured to perform a rule-based filtering step using time-domain and frequency-domain filtering, a spatial feature extraction step using a first neural network, and a temporal dynamics step using a second neural network. In this example or any other example, the flicker detection system additionally outputs a plurality of flicker parameters defining characteristics of the display flicker condition, and wherein the calibration parameter updates applied by the calibration adjustment system are determined based at least in part on the plurality of flicker parameters. In this example or any other example, the instructions are further executable to receive user input indicating whether the display flicker condition has improved, and wherein the reinforcement learning algorithm is further rewarded based at least in part on the user input indicating that the display flicker condition has improved. In this example or any other example, the flicker detection subsystem is previously trained based at least in part on a plurality of training image frames, along with corresponding training labels indicating whether the training image frames exhibit display flicker.
[0065] In an example, a method for flicker mitigation comprises: at a computing device, receiving a plurality of image frames depicting a digital display, wherein the digital display is affected by an ongoing display flicker condition; inputting the plurality of image frames to a flicker detection subsystem configured to output a severity indication that indicates an estimated severity of the display flicker condition, wherein the flicker detection subsystem is configured to perform a rule-based filtering step using time-domain and frequency-domain filtering, a spatial feature extraction step using a first neural network, and a temporal dynamics step using a second neural network; inputting the severity indication to a calibration adjustment subsystem configured to iteratively update an existing calibration policy defined by one or more calibration parameters of the digital display, wherein the calibration adjustment system implements a reinforcement learning algorithm that is rewarded for calibration parameter updates that reduce the severity indication; and based at least in part on determining that the calibration parameter updates have reduced the severity indication to below a threshold, replacing the existing calibration policy in firmware of the display device. In this example or any other example, the flicker detection system additionally outputs a plurality of flicker parameters defining characteristics of the display flicker condition, and wherein the calibration parameter updates applied by the calibration adjustment system are determined based at least in part on the plurality of flicker parameters.
[0066] It will be understood that the configurations and / or approaches described herein are exemplary in nature, and that these specific embodiments or examples are not to be considered in a limiting sense, because numerous variations are possible. The specific routines or methods described herein may represent one or more of any number of processing strategies. As such, various acts illustrated and / or described may be performed in the sequence illustrated and / or described, in other sequences, in parallel, or omitted. Likewise, the order of the above-described processes may be changed.
[0067] The subject matter of the present disclosure includes all novel and non-obvious combinations and sub-combinations of the various processes, systems and configurations, and other features, functions, acts, and / or properties disclosed herein, as well as any and all equivalents thereof.
Claims
1.A method for flicker mitigation, the method comprising:at a computing device, receiving a plurality of image frames depicting a digital display, wherein the digital display is affected by an ongoing display flicker condition;inputting the plurality of image frames to a flicker detection subsystem configured to output a severity indication that indicates an estimated severity of the display flicker condition;inputting the severity indication to a calibration adjustment subsystem configured to iteratively update an existing calibration policy defined by one or more calibration parameters of the digital display, wherein the calibration adjustment system implements a reinforcement learning algorithm that is rewarded for calibration parameter updates that reduce the severity indication; andbased at least in part on determining that the calibration parameter updates have reduced the severity indication to below a threshold, outputting an updated calibration policy based at least in part on the calibration parameter updates.2.The method of claim 1, wherein the flicker detection subsystem is configured to perform a rule-based filtering step using time-domain and frequency-domain filtering, a spatial feature extraction step using a first neural network, and a temporal dynamics step using a second neural network.3.The method of claim 2, wherein the first neural network is a convolutional neural network (CNN) , and the second neural network is a long short-term memory (LSTM) network.4.The method of claim 1, wherein the flicker detection system additionally outputs a plurality of flicker parameters defining characteristics of the display flicker condition, and wherein the calibration parameter updates applied by the calibration adjustment system are determined based at least in part on the plurality of flicker parameters.5.The method of claim 4, wherein the plurality of flicker parameters include at least one of a frequency of the display flicker condition, an amplitude of the display flicker condition, a periodicity of the display flicker condition, and localized brightness variations detected on the digital display.6.The method of claim 1, further comprising receiving user input indicating whether the display flicker condition has improved, and wherein the reinforcement learning algorithm is further rewarded based at least in part on the user input indicating that the display flicker condition has improved.7.The method of claim 1, wherein the flicker detection subsystem is previously trained based at least in part on a plurality of training image frames, along with corresponding training labels indicating whether the training image frames exhibit display flicker.8.The method of claim 1, wherein the digital display is integrated into a same physical housing as the computing device.9.The method of claim 8, wherein the plurality of image frames are captured by a camera system that is also integrated into the same physical housing as the computing device.10.The method of claim 1, wherein the plurality of image frames are captured by an external camera system communicatively coupled with the computing device, wherein the external camera system is in a separate physical housing from the computing device.11.The method of claim 1, wherein outputting the updated calibration policy includes replacing the existing calibration policy in firmware of the display device.12.The method of claim 1, wherein the one or more calibration parameters of the digital display include at least one of a pulse-width modulation (PWM) frequency of the digital display, a duty cycle of the digital display, a refresh rate of the digital display, and a phase compensation of the digital display.13.The method of claim 1, wherein the reinforcement learning algorithm is one of a deep-Q network (DQN) or a proximal policy optimization (PPO) algorithm.14.A computing device, comprising:a logic subsystem; anda storage subsystem holding instructions executable by the logic subsystem to:receive a plurality of image frames depicting a digital display, wherein the digital display is affected by an ongoing display flicker condition;input the plurality of image frames to a flicker detection subsystem configured to output a severity indication that indicates an estimated severity of the display flicker condition;input the severity indication to a calibration adjustment subsystem configured to iteratively update an existing calibration policy defined by one or more calibration parameters of the digital display, wherein the calibration adjustment system implements a reinforcement learning algorithm that is rewarded for calibration parameter updates that reduce the severity indication; andbased at least in part on determining that the calibration parameter updates have reduced the severity indication to below a threshold, output an updated calibration policy based at least in part on the calibration parameter updates.15.The computing device of claim 14, wherein the flicker detection subsystem is configured to perform a rule-based filtering step using time-domain and frequency-domain filtering, a spatial feature extraction step using a first neural network, and a temporal dynamics step using a second neural network.16.The computing device of claim 14, wherein the flicker detection system additionally outputs a plurality of flicker parameters defining characteristics of the display flicker condition, and wherein the calibration parameter updates applied by the calibration adjustment system are determined based at least in part on the plurality of flicker parameters.17.The computing device of claim 14, wherein the instructions are further executable to receive user input indicating whether the display flicker condition has improved, and wherein the reinforcement learning algorithm is further rewarded based at least in part on the user input indicating that the display flicker condition has improved.18.The computing device of claim 14, wherein the flicker detection subsystem is previously trained based at least in part on a plurality of training image frames, along with corresponding training labels indicating whether the training image frames exhibit display flicker.19.A method for flicker mitigation, the method comprising:at a computing device, receiving a plurality of image frames depicting a digital display, wherein the digital display is affected by an ongoing display flicker condition;inputting the plurality of image frames to a flicker detection subsystem configured to output a severity indication that indicates an estimated severity of the display flicker condition, wherein the flicker detection subsystem is configured to perform a rule-based filtering step using time-domain and frequency-domain filtering, a spatial feature extraction step using a first neural network, and a temporal dynamics step using a second neural network;inputting the severity indication to a calibration adjustment subsystem configured to iteratively update an existing calibration policy defined by one or more calibration parameters of the digital display, wherein the calibration adjustment system implements a reinforcement learning algorithm that is rewarded for calibration parameter updates that reduce the severity indication; andbased at least in part on determining that the calibration parameter updates have reduced the severity indication to below a threshold, replacing the existing calibration policy in firmware of the display device.20.The method of claim 19, wherein the flicker detection system additionally outputs a plurality of flicker parameters defining characteristics of the display flicker condition, and wherein the calibration parameter updates applied by the calibration adjustment system are determined based at least in part on the plurality of flicker parameters.
