Method and device for detecting bad channels of a probe, computer device and storage medium

Abstract

The embodiment of the present specification provides a detector bad channel detection method and device, computer equipment and storage medium, and relates to the technical field of computer processing. The method comprises: acquiring current projection data; the projection data is collected based on multiple channels by the detector; spatial feature analysis is performed on the current projection data to obtain first bad channel information of each channel; time feature analysis is performed on the current projection data to obtain second bad channel information of each channel; based on the first bad channel information and the second bad channel information, the comprehensive bad channel information of each channel is determined; and based on the comprehensive bad channel information, the bad channel of each channel is identified to obtain a bad channel identification result. In this way, the limitation of single-dimensional analysis can be broken through, real-time dynamic evaluation of each channel can be realized, and therefore the accuracy and robustness of detector channel detection can be improved.

Description

Technical Field

[0001] The embodiments described in this specification relate to the field of computer technology, specifically to a method, apparatus, computer device, and storage medium for detecting faulty channels in a detector. Background Technology

[0002] Multi-slice CT scanners typically contain dozens or even more channels. In actual operation, due to factors such as temperature, humidity, voltage fluctuations, mechanical vibration, or component aging, some channels of the CT scanner may experience performance degradation or malfunction, becoming "bad channels." These bad channels exhibit significantly different X-ray responses compared to normal channels, leading to various artifacts in the reconstructed CT images. For example, if a channel responds abnormally at all projection angles, it will appear as a ring-shaped artifact on the image; if it only responds abnormally at a specific angle, it will appear as a strip-shaped artifact. These artifacts severely degrade image quality and affect the accuracy of clinical diagnosis by physicians.

[0003] Therefore, it is necessary to provide a method for detecting faulty channels in a detector that can overcome the limitations of single-dimensional analysis and achieve real-time dynamic evaluation of each channel, thereby improving the accuracy and robustness of detector channel detection. Summary of the Invention

[0004] In view of this, various embodiments of this specification aim to provide a method, apparatus, computer device, and storage medium for detecting faulty channels in a detector, so as to improve the accuracy and robustness of detector channel detection.

[0005] This specification provides a method for detecting faulty channels in a detector. The method includes: acquiring current projection data; wherein the projection data is collected by the detector based on multiple channels; performing spatial feature analysis on the current projection data to obtain first faulty channel information for each of the channels; wherein the spatial feature analysis is based on the spatial distribution differences of the projection data corresponding to the channels; performing temporal feature analysis on the current projection data to obtain second faulty channel information for each of the channels; wherein the temporal feature analysis is based on the temporal distribution differences of the projection data corresponding to the channels; determining comprehensive faulty channel information for each of the channels based on the first and second faulty channel information; and identifying faulty channels for each of the channels based on the comprehensive faulty channel information to obtain a faulty channel identification result.

[0006] In some implementations, the step of performing spatial feature analysis on the current projection data to obtain the first bad channel information for each of the channels includes: inputting the current projection data into a pre-trained first analysis model to obtain the first bad channel information for each of the corresponding channels, wherein the first analysis model is trained using a first training dataset containing first projection data samples of bad channels and their corresponding bad channel mask labels; or, performing edge detection on the current projection data to obtain the first bad channel information for each channel; or, performing dynamic threshold analysis on the current projection data to obtain the first bad channel information for each channel; or, performing wavelet transform analysis on the current projection data to obtain the first bad channel information for each channel.

[0007] In some implementations, the current projection data is the projection data corresponding to the current laying-out View; the step of performing time feature analysis on the current projection data to obtain the second bad channel information of each of the channels includes: obtaining the time fluctuation index of each of the channels within a set time window; wherein, the set time window is a continuous time window before the current moment including a specified number of laying-out Views, and the time fluctuation index is used to characterize the volatility or dispersion of the projection data within the corresponding time window; and determining the second bad channel information of each of the channels based on the time fluctuation index.

[0008] In some implementations, the set time window is determined in the following manner: upon receiving a time window adjustment instruction for the set time window, the set time window is determined based on the time length information carried by the time window adjustment instruction; otherwise, the set time window is determined based on preset specified time length information.

[0009] In some implementations, the time fluctuation index is determined by the following formula: ; in, This represents the time fluctuation index at the current moment. This indicates that a time window is set. This represents the projection data of the channel at time i. and These represent the mean and standard deviation of the channel in the historical projection data for the same period, respectively. This is a normalization factor used to reduce the impact of a set time window.

[0010] In some implementations, the current projection data is the projection data corresponding to the current laying view; the step of performing time feature analysis on the current projection data to obtain the second bad channel information of each of the channels includes: inputting the current projection data and the projection data that are consecutive within a set time window before the current projection data into a pre-trained second analysis model to obtain the second bad channel information of each of the channels; wherein, the set time window is a time window that is consecutive before the current moment and includes a specified number of laying views; wherein, the second analysis model is trained using a second training dataset of second projection data samples from multiple consecutive historical moments and their corresponding time fluctuation index labels, and the time fluctuation index labels are determined based on the second projection data samples from multiple consecutive historical moments.

[0011] In some implementations, the step of identifying bad channels based on the comprehensive bad channel information to obtain a bad channel identification result includes: for any given channel, if the comprehensive bad channel information meets a set condition, identifying the channel as the bad channel, and obtaining a bad channel identification result including the bad channel; wherein, the comprehensive bad channel information of the channel is determined by the following formula: ; in, This indicates comprehensive information about bad channels. and These represent the information from the first bad channel and the second bad channel, respectively. The first weight parameter corresponds to the first bad channel information. The second weight parameter corresponds to the second bad channel information, and .

[0012] In some embodiments, the method further includes: in response to a weight adjustment instruction for the first weight parameter and the second weight parameter, redetermining the bad channel comprehensive information of each of the channels according to the adjusted first weight parameter and the second weight parameter; and / or, in response to a condition adjustment instruction for the set conditions, adjusting the set conditions, and re-identifying bad channels for each of the channels according to the adjusted set conditions to obtain an adjusted bad channel identification result.

[0013] In some embodiments, the method further includes: providing a user interface displaying channel identifiers corresponding to each of the channels and their comprehensive bad channel information; displaying a first marker corresponding to each of the channels on the user interface; wherein the first marker is determined based on the comprehensive bad channel information, and different first markers correspond to different comprehensive bad channel information; and displaying a second marker on the user interface; wherein the second marker indicates that the corresponding channel is a bad channel in the bad channel identification result.

[0014] In some embodiments, the detection method further includes: adjusting the bad channel identification result in response to a second mark adjustment operation on the user interface for any of the channels; wherein the second mark adjustment operation includes a second mark triggering operation and / or a second mark cancellation operation, the second mark triggering operation being used to identify any of the channels as the bad channel, and the second mark cancellation operation being used to re-identify any of the bad channels in the bad channel identification result as a normal channel.

[0015] This specification provides a detection device for faulty channels of a detector. The device includes: a projection data acquisition module for acquiring current projection data, wherein the projection data is collected by the detector based on multiple channels; a first feature analysis module for performing spatial feature analysis on the current projection data to obtain first faulty channel information for each of the channels, wherein the spatial feature analysis is based on the spatial distribution differences of the projection data corresponding to the channels; a second feature analysis module for performing temporal feature analysis on the current projection data to obtain second faulty channel information for each of the channels, wherein the temporal feature analysis is based on the temporal distribution differences of the projection data corresponding to the channels; a global evaluation module for determining comprehensive faulty channel information for each of the channels based on the first and second faulty channel information; and a channel identification module for identifying faulty channels for each of the channels based on the comprehensive faulty channel information to obtain a faulty channel identification result.

[0016] This specification provides a computer device including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the detection method described in any of the above embodiments.

[0017] This specification provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the detection method described in any of the above embodiments.

[0018] In the various embodiments provided in this specification, by acquiring current projection data collected by the detector based on multiple channels, spatial feature analysis is performed on the spatial distribution differences of each channel in the current projection data to obtain the first bad channel information for each channel. Similarly, temporal feature analysis is performed on the temporal distribution differences of each channel in the current projection data to obtain the second bad channel information for each channel. Then, based on the first and second bad channel information, the comprehensive bad channel information for each channel is determined, and bad channel identification is performed on each channel based on this comprehensive information to obtain the bad channel identification results. Channel masking is then performed based on these results. This approach overcomes the limitations of single-dimensional analysis, enabling real-time dynamic evaluation of each channel, thereby improving the accuracy and robustness of the detector's channel detection. Attached Figure Description

[0019] Figure 1 A schematic diagram of a detector fault channel detection system provided for embodiments of this specification; Figure 2 A flowchart illustrating the method for detecting a faulty channel in a detector as provided in the embodiments of this specification; Figure 3 A schematic diagram of a detector fault channel detection device provided for embodiments of this specification; Figure 4 A schematic diagram of a computer device provided for an embodiment of this specification. Detailed Implementation

[0020] To enable those skilled in the art to better understand the solutions described in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this specification, and not all of them. Based on the embodiments in this specification, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this specification.

[0021] In some related technologies, due to factors such as temperature, humidity, voltage fluctuations, mechanical vibration, or component aging, the channels of some CT detectors may experience performance degradation or malfunction, becoming unreliable or unstable bad channels. This can easily lead to various artifacts such as rings or stripes in the reconstructed CT images, severely reducing image quality and affecting the accuracy of doctors' clinical diagnoses.

[0022] This specification provides a scenario example of a method for detecting a faulty channel in a detector. This scenario example can be a detection system for a faulty channel in a detector. Please refer to [link / reference]. Figure 1 , Figure 1This is a schematic diagram of a detector fault channel detection system provided in the embodiments of this specification. The detection system may include a scanning device 110 and a terminal 120. The terminal 120 may be a computer or PC with image processing capabilities; for example, the terminal 120 may be, but is not limited to, a personal computer, laptop computer, tablet computer, etc. The scanning device 110 may include medical scanning equipment capable of CT scanning, conventional X-ray scanning equipment, or positron emission tomography (PET) equipment, etc., and the scanning device 110 can acquire projection data of the patient or subject at multiple times and multiple rotation angles.

[0023] For example, terminal 120 can be communicatively connected to scanning device 110 so that terminal 120 can acquire projection data collected by scanning device 110. This communication connection can be wired or wireless.

[0024] For example, the application scenario may also include a server 130, which can be communicatively connected to the terminal 120 and the scanning device 110 respectively, so that the server 130 can obtain the projection data sent by the scanning device 110 and send the obtained projection data to the terminal 120.

[0025] For example, taking a scanning device 110 that includes a CT detector as an example, the scanning device 110 may include an X-ray tube and a CT detector. During the scanning process of the patient or subject by the scanning device 110, the X-ray tube and the detector can rotate around it to acquire projection data at multiple rotation angles. Specifically, X-rays can be emitted through the X-ray tube and attenuated X-rays can be detected by the CT detector, thereby acquiring scanning data or projection data at various rotation angles during the rotation process.

[0026] In this scenario example, during the real-time acquisition of projection data by the CT detector, multi-dimensional feature analysis can be performed on the projection data. This approach avoids relying on a single dimension, such as spatial domain information, and comprehensively utilizes the spatial and temporal information features of the channel to identify bad channels, improving the reliability and robustness of the detection. Furthermore, by configuring an effective human-computer interaction mechanism, the operator's professional experience can be incorporated to confirm, adjust, and optimize the automated detection results, enhancing the intelligence and practicality of the detection process. This also ensures accuracy while meeting the streaming processing requirements of real-time scanning and image reconstruction. Thus, a CT bad channel detection method that integrates multi-dimensional information, operates dynamically in real-time, and allows for user interaction can improve the accuracy, adaptability, and clinical efficiency of bad channel detection.

[0027] This specification provides a method for detecting a faulty channel in a detector. Please refer to [link / reference]. Figure 2 , Figure 2 This is a flowchart illustrating a method for detecting faulty channels in a detector according to this embodiment. This embodiment provides the method's operational steps as shown in the flowchart, but based on conventional or non-inventive methods, more or fewer operational steps may be included. The order of steps listed in this embodiment is merely one possible execution order among many, and does not represent the only possible execution order. In actual systems or server products, the method can be executed sequentially as shown in the embodiment or in parallel (e.g., in a parallel processor or multi-threaded processing environment). This method for detecting faulty channels in a CT scanner can be applied to a terminal in a detection system, specifically as follows... Figure 2 As shown, the detection method may include the following steps.

[0028] Step S210: Obtain the current projection data; wherein, the projection data is collected by the detector based on multiple channels.

[0029] The following explanation uses a CT detector as an example. In some cases, multiple channels in a CT detector may include some faulty channels. Faulty channels, also known as unstable or unreliable detection channels, may acquire inaccurate or distorted projection data, resulting in image artifacts. Therefore, it is necessary to detect faulty channels in a timely manner to shield them and improve the accuracy and reliability of projection data acquisition.

[0030] The current projection data can refer to the projection data currently acquired by multiple channels of the CT detector, or it can refer to the set of raw signals acquired at the current moment.

[0031] For example, in a CT detector, each channel corresponds to a detection unit, and multiple channels are arranged in a preset array to receive X-rays passing through the scanned object and convert them into electrical signals to obtain corresponding projection values. For instance, in a multi-row CT detector with 64 channels, the current projection data contains 64 projection values ​​that correspond one-to-one with each channel.

[0032] Step S220: Perform spatial feature analysis on the current projection data to obtain the first bad channel information for each channel; wherein, the spatial feature analysis is based on the spatial distribution differences of the projection data corresponding to the channels.

[0033] The first bad channel information refers to channel information or channel evaluation information obtained from spatial feature analysis of the current projection data. It characterizes the tendency of each channel to become a bad channel, i.e., the instability of each channel. The first bad channel information can be the first bad channel probability, the first bad channel score, or the expected value of the first bad channel. The expected value of the first bad channel can be the product of the first bad channel probability and the first bad channel score. The first bad channel probability, the first bad channel score, or the expected value of the first bad channel refer to the bad channel probability, bad channel score, and bad channel expected value of each channel obtained from spatial feature analysis of the current projection data.

[0034] For example, the first bad channel information may include the first bad channel probability. The first bad channel probability may refer to the bad channel probability of each channel obtained by performing spatial feature analysis on the current projection data, which is used to characterize the possibility that a channel is judged as a bad channel due to abnormal spatial dimensional behavior.

[0035] Specifically, spatial feature analysis can refer to statistical or feature extraction operations based on the spatial distribution differences of each channel in the projection data. For example, spatial feature analysis can be achieved by identifying abnormal patterns in the spatial distribution of the projection values ​​or response values ​​corresponding to a channel. This can be done by detecting abrupt changes, statistical outliers, or frequency domain singularities to determine the likelihood that the channel is a bad channel. In this way, it is possible to capture channel anomaly information caused by instantaneous hardware damage or inherent defects from the spatial dimension and record it as the first bad channel information.

[0036] Step S230: Perform time feature analysis on the current projection data to obtain the second bad channel information for each channel; wherein, the time feature analysis is based on the time distribution differences of the projection data corresponding to the channel.

[0037] The second bad channel information refers to channel information or channel evaluation information obtained from temporal feature analysis of the current projection data. It characterizes the tendency of each channel to become a bad channel, i.e., the instability of each channel. The second bad channel information can be the second bad channel probability, second bad channel score, or expected value of the second bad channel. The expected value of the second bad channel can be the product of the second bad channel probability and the second bad channel score. The second bad channel probability, second bad channel score, or expected value of the second bad channel refer to the bad channel probability, bad channel score, and expected value of each channel obtained from temporal feature analysis of the current projection data.

[0038] For example, the second bad channel information may include the second bad channel probability. The second bad channel probability may refer to the bad channel probability of each channel obtained by performing temporal feature analysis on the current projection data, and is used to characterize the possibility that a channel is judged as a bad channel due to abnormal behavior in the temporal dimension.

[0039] Specifically, time feature analysis can refer to a dynamic evaluation based on the distribution differences of any channel in the projection data over a time series. For example, this can be achieved by evaluating the fluctuation characteristics of the projection value or response value of any channel within a continuous scanning window. Here, the scanning window refers to the scanning time window corresponding to the scanning process.

[0040] As an example, for any channel, the time fluctuation index within a set time window corresponding to the current moment can be calculated, and this time fluctuation index can be compared with the historical fluctuation baseline of that channel to determine the bad channel information of that channel, which is then recorded as the second bad channel information. Here, the set time window can refer to a pre-set time window used to determine the time fluctuation index. For example, the set time window can be the time window corresponding to a specified number of X-ray views. A X-ray view refers to the process of X-rays emitted from the X-ray tube reaching the CT detector.

[0041] For example, the time window can be the time window corresponding to 50 line-laying views.

[0042] For example, for any channel, the time fluctuation index of the channel in the most recent 50 consecutive line views can be calculated. If the volatility represented by the current time fluctuation index is higher than the channel's historical level for the same period, then higher second bad channel information can be assigned. In this way, channel performance degradation caused by component aging, voltage instability or intermittent faults can be captured from the time dimension, supplementing time-domain faults that are difficult to detect by spatial feature analysis and improving the timing sensitivity of detection.

[0043] Step S240: Based on the first bad channel information and the second bad channel information, determine the comprehensive bad channel information for each channel.

[0044] Specifically, the comprehensive bad channel information refers to the evaluation information obtained by comprehensively assessing the first bad channel information based on the spatial dimension and the second bad channel information based on the temporal dimension. It can comprehensively describe the bad channel condition or instability of the corresponding channel in both the temporal and spatial dimensions. The comprehensive bad channel information can be bad channel confidence, bad channel comprehensive score, or bad channel comprehensive expected value, etc.

[0045] For example, comprehensive information on bad channels can include the confidence level of the bad channel. The confidence level of the bad channel can be used to indicate the degree of trust in determining whether a corresponding channel is a bad channel.

[0046] Specifically, taking the information of the first bad channel, including the probability of the first bad channel, the information of the second bad channel, including the probability of the second bad channel, and the comprehensive information of the bad channel, including the confidence level of the bad channel, as an example, for any channel, the confidence level of the bad channel can be determined comprehensively based on the probability of the first bad channel and the probability of the second bad channel corresponding to that channel. In this way, the channel analysis results in both spatial and temporal dimensions can be integrated to form a final quantitative assessment of the channel status. By integrating spatial static features and temporal dynamic features, a more comprehensive and reliable assessment of the channel status can be formed, reducing the risk of false detection or false negative detection that may be caused by single-dimensional analysis and improving the overall robustness of bad channel determination.

[0047] Step S250: Based on the comprehensive information of bad channels, identify bad channels for each channel to obtain bad channel identification results, and then perform channel masking based on the bad channel identification results.

[0048] Among them, bad channel identification can refer to detecting the situation where the corresponding channel is a bad channel based on the comprehensive information of bad channels.

[0049] As an example, bad channel identification can refer to the decision-making process of determining whether a channel is a bad channel based on comprehensive information about the bad channel, such as the numerical value of the bad channel confidence score.

[0050] For example, the bad channel identification result may or may not include bad channels. If the bad channel identification result includes bad channels, it indicates that there are unstable bad channels among the multiple channels in the CT detector. If the bad channel identification result does not include bad channels, it indicates that no bad channels were identified among the multiple channels in the CT detector.

[0051] The results of bad channel identification can be used for subsequent channel masking. For example, before image reconstruction, the bad channel data marked in the bad channel identification results can be removed or interpolated for repair. In this way, the quantitative confidence assessment can be transformed into a clear channel state classification, generating execution instructions that can be directly used to guide image quality optimization. This ensures that channel masking based on bad channel identification results can effectively improve the quality of the final reconstructed image and reduce the generation of image artifacts.

[0052] In the above implementation, feature analysis is performed by integrating spatial and temporal dimensions, and decision-making is made using the comprehensive calculated confidence of bad channels. This can enhance the accuracy and reliability of the assessment, reduce the risk of false positives and false negatives, effectively capture inherent hardware problems, improve the accuracy and robustness of CT bad channel detection, provide a higher quality data foundation for image reconstruction, thereby suppressing artifact generation at the source, and flexibly adapt to different CT detectors and scanning scenarios to meet the real-time requirements of clinical streaming data processing.

[0053] In some implementations, spatial feature analysis can be based on a neural network model.

[0054] In this embodiment, spatial feature analysis of the current projection data to obtain the first bad channel information for each channel may include: inputting the current projection data into a pre-trained first analysis model for information prediction to obtain the first bad channel information for each corresponding channel. The first analysis model may be trained using a first training dataset containing first projection data samples of bad channels and their corresponding bad channel mask labels.

[0055] Wherein, the first projection data sample is the projection data sample used to train the first analysis model, and the first training dataset is the training dataset used to train the first analysis model.

[0056] For example, the first training dataset can contain a large number of projected data samples with known bad channel locations. Each projected data sample is labeled with a bad channel mask to identify the bad channel location. These bad channel mask labels serve as ground truth values ​​to guide model training in supervised learning. In this way, the limitations of traditional methods that rely on human experience to set features and thresholds can be avoided, improving the automation and accuracy of spatial feature analysis.

[0057] As an example, the first analysis model can be an end-to-end neural network model. That is, based on the first analysis model, the required first bad channel information can be directly mapped from the original projection data. For example, the end-to-end neural network model can employ a deep convolutional neural network structure, or network architectures such as ResNet and DenseNet. These network structures have hierarchical feature extraction capabilities, enabling them to learn various spatial features from the projection data. Inputting the current projection data into the pre-trained first analysis model for information prediction means using the entire current projection data as model input. Through multi-layer network computation within the first analysis model, the final output is the first bad channel information corresponding to each channel. This achieves efficient and accurate spatial feature analysis, improving the computational efficiency of the detection process while ensuring the accuracy of CT bad channel detection, thus meeting the needs of real-time processing.

[0058] In some implementations, spatial feature analysis can also be performed using methods such as edge detection, dynamic thresholding, or wavelet transform. That is, performing spatial feature analysis on the current projection data to obtain the first bad channel information for each channel can include: analyzing the current projection data based on methods such as edge detection, dynamic thresholding, or wavelet transform to obtain the first bad channel information for each channel. Alternatively, performing spatial feature analysis on the current projection data can refer to performing edge detection, dynamic thresholding analysis, or wavelet transform analysis on the current projection data to obtain the first bad channel information for each channel.

[0059] In some implementations, analyzing the current projected data based on edge detection to obtain the first bad channel information for each channel may include the following steps: Representing the response values ​​of each channel in the current projected data as one-dimensional discrete signals; where the response value is also the projected value. Calculating the gradient operator for each one-dimensional discrete signal, where the gradient operator is the first or second derivative of the one-dimensional discrete signal. Taking the absolute value of each gradient operator to obtain the corresponding gradient magnitude; a larger gradient magnitude indicates a greater difference between the corresponding data point and surrounding data points, more like abrupt change points or edge points, where the data point is also the projected value. Normalizing and mapping information based on each gradient magnitude to obtain the first bad channel information for each channel.

[0060] Specifically, taking the first bad channel information as the first bad channel probability as an example, normalization and probability mapping can refer to: normalizing the calculated gradient magnitude sequence including each gradient magnitude so that its range is between [0, 1], thereby obtaining the first bad channel probability of each channel.

[0061] For example, normalization and probability mapping can be determined by Equation 1.

[0062] P_error_edge[i]=(G[i]- min(G)) / (max(G)-min(G)) Formula 1; Where i represents the projection value of any channel, G[i] represents the gradient magnitude corresponding to the projection value of any channel, max(G) represents the maximum value in the gradient magnitude sequence, min(G) represents the minimum value in the gradient magnitude sequence, and P_error_edge[i] represents the probability of the first bad channel for any channel.

[0063] In some implementations, analyzing the current projection data based on a dynamic threshold to obtain the probability of the first bad channel for each channel may include the following steps: setting a sliding data window; calculating dynamic statistics; setting a dynamically adjustable probability threshold range; calculating the degree of deviation; and mapping it to the probability of the first bad channel.

[0064] For example, setting a sliding data window can be done by first setting the size W of the data window. Specifically, the response values ​​of each channel in the current projected data can be recorded as corresponding data points. For any given data point, the preceding W data points are taken as the reference data window. For example, dynamic statistics can include the moving average and the moving standard deviation. The dynamic statistics can be calculated within the reference data window by calculating the moving average and the moving standard deviation. For example, setting a probability threshold range can be done based on the calculated moving average and the moving standard deviation. For example, calculating the degree of deviation can be done by calculating the degree to which the current data point deviates from the center value of the probability threshold range, thus obtaining the corresponding deviation value. For example, mapping to the probability of the first bad channel can be done by normalizing the deviation value and mapping it to the [0, 1] interval, thus obtaining the probability of the first bad channel.

[0065] In some implementations, the current projection data can be the projection data corresponding to the current line-laying View.

[0066] Taking the second-bad-channel information, including the second-bad-channel probability, as an example, this embodiment performs time feature analysis on the current projection data to obtain the second-bad-channel information for each channel, that is, to obtain the second-bad-channel probability for each channel. The specific process may include: obtaining the time fluctuation index of each channel within a set time window; and determining the second-bad-channel probability for each channel based on the time fluctuation index. The set time window is a continuous time window preceding the current moment, including a specified number of laid-out views. The time fluctuation index is used to characterize the volatility or dispersion of the projection data within the corresponding time window. It should be noted that the second-bad-channel probability for each channel can be determined directly based on the time fluctuation index, or it can be determined based on the time fluctuation index combined with methods such as neural network models.

[0067] Specifically, to improve the flexibility and accuracy of bad channel detection, the time window is adjustable. The time window can be determined as follows: upon receiving a time window adjustment command, the time window is determined based on the time length information carried in the command; otherwise, the time window is determined based on a preset specified time length. In this way, users can flexibly adjust the time window according to the actual detection scenario. As an example, the time window adjustment command and its carried time length information can be determined based on the user interface.

[0068] Specifically, historical projection data for each channel can also be cached. The time fluctuation index is determined using Formula 2.

[0069] Formula 2; in, This represents the time fluctuation index at the current moment. This indicates that a time window is set. This represents the projection data, projection value, or response value of the channel at time i. and These represent the mean and standard deviation of the channel's projection data over the same historical period, where "same period" refers to the same rotation angle. This is a normalization factor used to reduce the impact of a set time window.

[0070] For example, after receiving the projection data corresponding to the current line-laying View each time and calculating the corresponding time fluctuation index, the mean and standard deviation of the historical projection data for the same period can be updated cyclically.

[0071] In the above implementation, the process of determining the time fluctuation index at the current moment based on the current projection data involves calculating the absolute deviation of each data point from its long-term benchmark / mean value and performing standardization. This eliminates the influence of dimensions, allowing the determined time fluctuation index to be compared across different data points / projection data. The larger the time fluctuation index, the stronger and more unstable the fluctuation of the projection data within the corresponding set time window, and the more likely the corresponding channel is a bad channel. Conversely, the smaller the time fluctuation index, the more stable the fluctuation of the projection, the closer it is to its historical normal, and the more likely the corresponding channel is a normal channel.

[0072] In some implementations, performing time feature analysis on the current projection data to obtain second-bad-channel information for each channel may include: inputting the current projection data and consecutive projection data within a set time window preceding the current projection data into a pre-trained second analysis model to obtain second-bad-channel information for each channel. The set time window can be a consecutive time window preceding the current moment, including a specified number of laying views, and includes the current moment.

[0073] The second analytical model can be trained using a second training dataset consisting of second projected data samples from multiple consecutive historical moments and their corresponding time fluctuation index labels. The time fluctuation index labels can be determined based on these second projected data samples from multiple consecutive historical moments.

[0074] The second projected data sample is the projected data sample used to train the second analysis model, and the second training dataset is the training dataset used to train the second analysis model. The time fluctuation index label can be determined using Formula 2, where T in Formula 2 represents the time length of the time window corresponding to the multiple consecutive historical moments.

[0075] For example, the second analysis model can be determined based on a multilayer perceptron (MLP), a long short-term memory (LSTM) network, or a recurrent neural network (RNN).

[0076] In some implementations, bad channel identification is performed on each channel based on bad channel comprehensive information to obtain bad channel identification results. This may include: for any channel, if the bad channel comprehensive information meets the set conditions, the channel is identified as a bad channel, and a bad channel identification result including bad channels is obtained.

[0077] For example, taking the comprehensive information of bad channels, including the confidence level of bad channels, as an example, a confidence threshold for identifying bad channels can be set. In the CT bad channel detection method, bad channel identification is performed on each channel based on the comprehensive information of bad channels to obtain bad channel identification results. This can include: for any channel, if the confidence level of bad channels is greater than or equal to the confidence threshold, the channel is determined to be a bad channel, thus obtaining a bad channel identification result including bad channels.

[0078] In some implementations, a first weight parameter and a second weight parameter can be set for the first bad channel information and the second bad channel information, respectively, and the comprehensive bad channel information of each channel can be determined based on the first bad channel information and its first weight parameter, the second bad channel information and its second weight parameter.

[0079] Specifically, the comprehensive information on bad channels can be determined using Formula 3.

[0080] Formula 3; in, This indicates comprehensive information about bad channels. and These represent the information from the first bad channel and the second bad channel, respectively. The first weight parameter corresponds to the first bad channel information. The second weight parameter corresponds to the second bad channel information, and .

[0081] For example, the comprehensive information on bad channels can be the confidence level of bad channels.

[0082] In some embodiments, the method for detecting bad channels in CT scans may further include: responding to a weight adjustment instruction for a first weight parameter and a second weight parameter, redetermining the comprehensive bad channel information for each channel based on the adjusted first weight parameter and second weight parameter; and / or, responding to an adjustment instruction for set conditions, adjusting the set conditions, and re-identifying bad channels for each channel based on the adjusted set conditions to obtain an adjusted bad channel identification result. This allows for adjustment of the identification sensitivity.

[0083] For example, the weight adjustment instruction may be determined based on the provided user interface or through other means.

[0084] In some implementations, the method for detecting bad channels in CT scans may further include: displaying a first marker corresponding to each channel on a user interface; wherein the first marker is determined based on comprehensive information about bad channels, and different first markers correspond to different comprehensive information about bad channels.

[0085] For example, the user interface displays the channel identifier and the confidence level of the bad channel for each channel; the user interface also displays a first marker corresponding to each channel.

[0086] The first label can be determined based on the confidence level of the bad channel, and different first labels correspond to different confidence levels of the bad channel.

[0087] For example, the first marker can be a color marker. As one example, the first marker can refer to the color intensity or shade of each channel identifier, determined by the bad channel confidence level of the corresponding channel. For instance, a channel with a higher bad channel confidence level will have its corresponding channel identifier displayed in a brighter color. As another example, the first marker can refer to the color intensity or shade of the bad channel confidence level displayed on the user interface for each channel.

[0088] In some implementations, the method for detecting bad channels in CT scans may further include: displaying a second marker on a user interface; wherein the second marker indicates that the corresponding channel is a bad channel in the bad channel identification results.

[0089] For example, the second marker can be a text marker, an image marker, and / or a symbol marker. For instance, it could display "Bad Channel," "Unstable Channel," "Wrong," "×," "!", etc., on a user interface.

[0090] In some implementations, the method for detecting bad channels in CT scans may further include: adjusting the bad channel identification result in response to a second marker adjustment operation for any channel on a user interface.

[0091] Specifically, the second marker adjustment operation may include a second marker trigger operation and / or a second marker cancellation operation.

[0092] The second flag triggering operation is used to identify any channel as a bad channel, and the second flag cancellation operation is used to re-identify any bad channel in the bad channel identification results as a normal channel.

[0093] In some implementations, the method for detecting bad channels in CT scans, after adjusting the bad channel identification result in response to the second label adjustment operation, may further include: adjusting and updating the parameters of the bad channel information model based on the adjusted bad channel identification result, so that the bad channel information model better meets user needs. The bad channel information model can be a first analysis model for spatial feature analysis or a second analysis model for temporal feature analysis.

[0094] Specifically, the adjusted bad channel identification results can be used as new bad channel mask labels, forming incremental training samples with the corresponding original projection data. The goal is to reduce the difference between the bad channel information output by the bad channel information model based on the original projection data and the new bad channel mask labels. The parameters of the bad channel information model are fine-tuned using the gradient descent algorithm to complete the model update.

[0095] This specification provides a device for detecting faulty channels in a detector. Please refer to [link / reference]. Figure 3 The CT bad channel detection device may include a projection data acquisition module 310, a first feature analysis module 320, a second feature analysis module 330, a global evaluation module 340, and a channel recognition module 350.

[0096] The projection data acquisition module 310 is used to acquire the current projection data; wherein, the projection data is collected by the detector based on multiple channels; The first feature analysis module 320 is used to perform spatial feature analysis on the current projection data to obtain the first bad channel information for each channel; wherein, the spatial feature analysis is based on the spatial distribution differences of the projection data corresponding to the channels; The second feature analysis module 330 is used to perform time feature analysis on the current projection data to obtain the second bad channel information for each channel; wherein, the time feature analysis is based on the time distribution differences of the projection data corresponding to the channel; The global evaluation module 340 is used to determine the comprehensive bad channel information of each channel based on the first bad channel information and the second bad channel information. The channel identification module 350 is used to identify bad channels based on comprehensive bad channel information, obtain bad channel identification results, and then perform channel masking based on the bad channel identification results.

[0097] The specific functions and effects of the detector fault channel detection device can be explained by referring to other embodiments in this specification, and will not be repeated here. Each module in the detector fault channel detection device can be implemented entirely or partially through software, hardware, or a combination thereof. Each module can be embedded in the processor of the computer device in hardware form or independent of it, or it can be stored in the memory of the computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0098] This specification also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a computer, implements the CT bad channel detection method in any of the above embodiments.

[0099] This specification also provides a computer program product containing instructions that, when executed by a computer, cause the computer to implement the CT bad channel detection method in any of the above embodiments.

[0100] This specification also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, which, when executed by the processor, implements the CT bad channel detection method in any of the above embodiments.

[0101] In some implementations, please refer to Figure 4 The computer device can be a terminal, and its internal structure diagram can be as follows: Figure 4 As shown. The computer device includes a processor, memory, and communication interface connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. When the computer program is executed by the processor, it implements a method for detecting bad channels in CT scans.

[0102] It is understood that the specific examples in this document are only intended to help those skilled in the art better understand the embodiments described herein, and are not intended to limit the scope of the invention.

[0103] It is understood that in the various embodiments described in this specification, the sequence number of each process does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments described in this specification.

[0104] It is understood that the various implementation methods described in this specification can be implemented individually or in combination, and the implementation methods in this specification are not limited in this respect.

[0105] Unless otherwise stated, all technical and scientific terms used in the embodiments of this specification have the same meaning as commonly understood by one of ordinary skill in the art. The terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the scope of this specification. The term "and / or" as used in this specification includes any and all combinations of one or more of the associated listed items. The singular forms "a," "the," and "the" as used in the embodiments of this specification and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise.

[0106] It is understood that the processor in the embodiments of this specification can be an integrated circuit chip with signal processing capabilities. In implementation, each step of the above method embodiments can be completed by integrated logic circuits in the processor's hardware or by instructions in software form. The processor can be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this specification. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this specification can be directly implemented by a hardware decoding processor, or by a combination of hardware and software modules in the decoding processor. The software modules can reside in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. This storage medium is located in memory; the processor reads information from the memory and, in conjunction with its hardware, completes the steps of the above methods.

[0107] It is understood that the memory in the embodiments of this specification may be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. Non-volatile memory may be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. Volatile memory may be random access memory (RAM). It should be noted that the memory in the systems and methods described herein is intended to include, but is not limited to, these and any other suitable types of memory.

[0108] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this specification.

[0109] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the aforementioned method implementations, and will not be repeated here.

[0110] The above description is merely a specific embodiment of this specification, but the scope of protection of this invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this specification should be included within the scope of protection of this specification. Therefore, the scope of protection of this invention should be determined by the scope of the claims.

Claims

1. A method for detecting a faulty channel in a detector, characterized in that, The method includes: Acquire the current projection data; where the projection data is collected by the detector based on multiple channels; Spatial feature analysis is performed on the current projection data to obtain the first bad channel information for each of the channels; wherein, the spatial feature analysis is based on the spatial distribution differences of the projection data corresponding to the channels; A time feature analysis is performed on the current projection data to obtain the second bad channel information for each of the channels; wherein, the time feature analysis is based on the time distribution differences of the projection data corresponding to the channels; Based on the first bad channel information and the second bad channel information, the comprehensive bad channel information of each channel is determined; Based on the comprehensive information of the bad channels, bad channel identification is performed on each of the channels to obtain the bad channel identification result.

2. The detection method according to claim 1, characterized in that, The step of performing spatial feature analysis on the current projection data to obtain the first bad channel information for each of the channels includes: The current projection data is input into a pre-trained first analysis model to obtain the first bad channel information for each of the corresponding channels. The first analysis model is trained using a first training dataset containing first projection data samples of bad channels and their corresponding bad channel mask labels. Alternatively... Perform edge detection on the current projection data to obtain the first bad channel information for each channel; or, Perform dynamic threshold analysis on the current projection data to obtain the first bad channel information for each channel; or, Wavelet transform analysis is performed on the current projection data to obtain the first bad channel information for each channel.

3. The detection method according to claim 1, characterized in that, The current projection data is the projection data corresponding to the current line-laying View; The step of performing time feature analysis on the current projection data to obtain the second bad channel information for each of the channels includes: Obtain the time fluctuation index of each of the channels within a set time window; wherein, the set time window is a continuous time window including a specified number of line views before the current time, and the time fluctuation index is used to characterize the volatility or dispersion of the projection data within the corresponding time window; Based on the time fluctuation index, the second bad channel information for each of the channels is determined.

4. The detection method according to claim 3, characterized in that, The set time window is determined in the following way: If a time window adjustment command for the set time window is received, the set time window is determined based on the time length information carried in the time window adjustment command; otherwise, The set time window is determined based on the preset specified time length information.

5. The detection method according to claim 3, characterized in that, The time fluctuation index is determined by the following formula: ； in, This represents the time fluctuation index at the current moment. This indicates that a time window is set. This represents the projection data of the channel at time i. and These represent the mean and standard deviation of the channel in the historical projection data for the same period, respectively. This is a normalization factor used to reduce the impact of a set time window.

6. The detection method according to claim 1, characterized in that, The current projection data is the projection data corresponding to the current line-laying View; The step of performing time feature analysis on the current projection data to obtain the second bad channel information for each of the channels includes: The current projection data and the projection data that occurred consecutively within a set time window before the current projection data are input into the pre-trained second analysis model to obtain the second bad channel information for each of the channels. The set time window is a continuous time window before the current moment, including a specified number of line-laying views; The second analysis model is trained using a second training dataset consisting of second projected data samples from multiple consecutive historical moments and their corresponding time fluctuation index labels. The time fluctuation index labels are determined based on the second projected data samples from the multiple consecutive historical moments.

7. The detection method according to claim 1, characterized in that, The step of identifying bad channels based on the comprehensive information of the bad channels to obtain bad channel identification results includes: For any of the aforementioned channels, if the comprehensive information of the bad channels meets the set conditions, the channel is identified as the bad channel, and a bad channel identification result including the bad channel is obtained; wherein, the comprehensive information of the bad channels is determined by the following formula: ； in, This indicates comprehensive information about bad channels. and These represent the information from the first bad channel and the second bad channel, respectively. The first weight parameter corresponds to the first bad channel information. The second weight parameter corresponds to the second bad channel information, and .

8. The detection method according to claim 7, characterized in that, The method further includes: In response to the weight adjustment instruction for the first weight parameter and the second weight parameter, the bad channel synthesis information for each of the channels is re-determined based on the adjusted first and second weight parameters; and / or, In response to the condition adjustment command for the set conditions, the set conditions are adjusted, and based on the adjusted set conditions, each of the channels is re-identified as a bad channel to obtain the adjusted bad channel identification result.

9. The detection method according to claim 1, characterized in that, The method further includes: A user interface is provided, which displays the channel identifiers corresponding to each of the channels and their comprehensive information on the bad channels. On the user interface, a first marker corresponding to each of the channels is displayed; wherein, the first marker is determined based on the comprehensive information of the bad channels, and the comprehensive information of the bad channels corresponding to different first markers is different; On the user interface, a second marker is displayed; wherein the second marker indicates that the corresponding channel is a bad channel in the bad channel identification result.

10. The detection method according to claim 9, characterized in that, The detection method further includes: In response to a second mark adjustment operation on the user interface for any of the channels, the bad channel identification result is adjusted; wherein, the second mark adjustment operation includes a second mark trigger operation and / or a second mark cancellation operation, the second mark trigger operation is used to identify any of the channels as bad channels, and the second mark cancellation operation is used to re-identify any of the bad channels in the bad channel identification result as normal channels.

11. A device for detecting faulty channels in a detector, characterized in that, The device includes: The projection data acquisition module is used to acquire the current projection data; the projection data is collected by the detector based on multiple channels. The first feature analysis module is used to perform spatial feature analysis on the current projection data to obtain the first bad channel information for each of the channels; wherein, the spatial feature analysis is based on the spatial distribution differences of the projection data corresponding to the channels; The second feature analysis module is used to perform time feature analysis on the current projection data to obtain the second bad channel information for each of the channels; wherein, the time feature analysis is based on the time distribution differences of the projection data corresponding to the channels; The global evaluation module is used to determine the comprehensive bad channel information of each of the channels based on the first bad channel information and the second bad channel information; The channel identification module is used to identify bad channels based on the comprehensive information of bad channels, and obtain bad channel identification results.

12. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the detector bad channel detection method according to any one of claims 1 to 10.

13. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the method for detecting bad channels of the detector as described in any one of claims 1 to 10.

Images