A critical part recognition method using region perception

Abstract

The application discloses a kind of to adopt regional perception's key part identification method, on the basis of preliminary completion target identification, using the obtained perception information, directly to the target area is carefully attention allocation;Subsequently, adaptive attention module dynamically adjusts its working gravity center, carries out depth analysis to the key part in target area, effectively captures and distinguishes the features of these parts, to realize high-precision key part identification.This method not only reduces unnecessary computing overhead, but also ensures the accuracy and robustness of key part identification.The present application optimizes the calculation process and intelligent resource allocation, provides an efficient target and key part identification scheme for embedded systems and other resource-constrained scenarios, and is expected to improve the intelligent level in the fields of intelligent monitoring, automated driving, medical image analysis, etc., and promote the intelligent development and application of the industry.

Description

Technical Field

[0001] The present invention belongs to the technical field of pattern recognition, and in particular relates to a method for identifying key parts using regional perception. Background Art

[0002] With the deepening evolution of digital transformation, the field of object detection and recognition is undergoing a revolutionary leap, moving from simple object recognition to precise positioning and deep structural analysis, marking a significant expansion of the technology's application scope. This progress not only broadens the technology's application prospects in fields such as security monitoring, industrial automation, and medical image analysis, but also heralds a new era of cross-industry collaborative innovation.

[0003] A range of advanced algorithms have demonstrated exceptional performance in object detection and classification in complex visual environments, laying a solid foundation for data processing pipelines. Current object part recognition technology integrates instance segmentation and semantic segmentation techniques to precisely delineate target entities with pixel-level accuracy, while assigning clear category labels to each pixel. This enables structural characterization, state assessment, and functional analysis of targets. Leveraging deep learning techniques, object part recognition technology can accurately identify key points or specific areas, enabling superior results in detecting subtle defects in industrial inspections and identifying and analyzing threat sources in intelligent security.

[0004] However, the challenges encountered in actual deployment cannot be ignored. In particular, high computing resource requirements, limited real-time processing capabilities, and interference from complex external factors have become key factors hindering the popularization and application of this technology. In a resource-constrained environment, the reliance on high-performance AI computing platforms not only significantly raises the hardware cost threshold but also directly hinders the widespread adoption of the technology. Furthermore, interference from complex factors, such as varying lighting conditions, severely impacts segmentation results, further highlighting the technology's fragility and limitations in real-world applications.

[0005] Disadvantages of existing technology:

[0006] (1) High computing resource consumption: Current target detection and recognition technologies face high computing resource requirements when deployed in practice, especially in resource-constrained environments such as embedded systems. This not only increases hardware costs but also limits the widespread application of the technology.

[0007] (2) Limited real-time processing capabilities: The technology is not capable of real-time response in processing high-dimensional data and dynamic environments, making it difficult to meet the needs of scenarios such as real-time monitoring and autonomous driving.

[0008] (3) Sensitive to complex external factors: When dealing with complex factors such as illumination changes, occlusion, and background clutter, the existing technology's segmentation effect and recognition accuracy are significantly reduced, exposing its fragility and limitations in real-world applications.

[0009] (4) Non-optimal resource allocation: Traditional methods usually perform calculations on the entire image and fail to intelligently focus on the target area, resulting in resource waste and low processing efficiency.

[0010] (5) Difficulty in identifying key parts: In complex scenarios, it is challenging to accurately identify and distinguish the key parts of a target, especially under conditions of limited computing resources. Summary of the Invention

[0011] In order to overcome the shortcomings of the existing technology, the present invention provides a method for identifying key parts using regional perception, aiming to achieve accurate identification of key parts of the target through efficient regional perception and adaptive attention mechanism, which is particularly suitable for environments with limited computing resources such as embedded systems. On the basis of preliminarily completing target identification, the present invention uses the obtained perception information to directly allocate detailed attention to the area where the target is located, rather than the entire image, significantly reducing computational complexity and improving processing efficiency. Through regional perception technology, the present invention can intelligently focus on the target area and avoid redundant processing of background information. Subsequently, the adaptive attention module dynamically adjusts its working focus, performs in-depth analysis of key parts within the target area, effectively captures and distinguishes the characteristics of these parts, thereby achieving high-precision key part identification. This method not only reduces unnecessary computational overhead, but also ensures the accuracy and robustness of key part identification. By optimizing the computational process and intelligent resource allocation, the present invention provides an efficient target and key part identification solution for embedded systems and other resource-constrained scenarios, which is expected to improve the intelligence level in fields such as intelligent monitoring, automated driving, and medical image analysis, and promote the intelligent development and application of the industry.

[0012] The technical solutions adopted by the present invention to solve the technical problems are as follows:

[0013] Step 1: Target detection and region preselection;

[0014] Use the target detection algorithm to perform global detection on the input image to locate all possible targets;

[0015] Output the bounding box coordinates of the target;

[0016] Step 2: Region perception and feature extraction;

[0017] Based on the results of target detection, focus on the area where each target is located; use the feature map generated during the detection network inference, combined with the detection results in step 1, to perform feature extraction and information fusion;

[0018] Step 3: Adaptive attention allocation;

[0019] Using an adaptive attention mechanism to dynamically adjust attention distribution, the key part recognition model focuses on the key parts of the target rather than the background or secondary areas;

[0020] Step 4: Identification and analysis of key parts;

[0021] Use the features filtered by the attention mechanism to identify and analyze key areas;

[0022] Step 5: Result output and optimization feedback;

[0023] The identified key parts information is integrated to form the final recognition result, and data from the execution process is collected for continuous optimization and parameter adjustment of the model.

[0024] Furthermore, the step 1 is specifically as follows:

[0025] Step 1-1: After obtaining the original or image-conditioned image from the image information source, a deep convolutional neural network is used to perform multi-layer convolution and pooling operations on the input image to extract feature maps at different levels, automatically learning the characteristic patterns related to the target in the image, and thus capturing the target feature information;

[0026] Step 1-2: Based on the prior knowledge of the target shape and size, by presetting a variety of different anchor boxes, the possibility of detecting targets of different sizes and shapes is increased, thereby improving the recall rate of detection;

[0027] Step 1-3: Use the non-maximum suppression algorithm to suppress bounding boxes with high overlap and low scores to ensure the accuracy of the final output bounding box.

[0028] Furthermore, the step 2 is specifically as follows:

[0029] Step 2-1: Based on the target area bounding box determined by target detection, combined with the feature map generated during the detection network inference, the corresponding area is processed using a lightweight feature extraction module;

[0030] Step 2-2: Through weighted summation, concatenation, and the use of attention mechanisms, features at different levels are fused according to certain rules to obtain a more comprehensive and representative feature representation;

[0031] Step 2-3: By sharing feature information with the detection network and realizing regional perception based on the detection results, key features are retained for subsequent identification of key parts of the target.

[0032] Furthermore, the step 3 is specifically as follows:

[0033] Step 3-1: Introduce a learnable weight matrix with the same dimension as the feature map. For the eigenvalue at each position, calculate its product with the element at the corresponding position in the weight matrix to obtain the weighted eigenvalue.

[0034] Assume that the input feature map is F∈R H×W×C , where H and W represent the height and width of the feature map respectively, C represents the number of channels, and the learnable weight matrix is ​​A∈R H×W×C ;

[0035] The calculation process of attention allocation is expressed as: F attended =F⊙A, where ⊙ represents element-wise multiplication operation.

[0036] Step 3-2: Normalize the weighted feature map.

[0037] Furthermore, the step 5 is specifically as follows:

[0038] Step 5-1: Organize key part information into a unified data structure, and perform data verification and exception handling before output;

[0039] Step 5-2: Collect various data during execution, including input, intermediate and final results, system running time and resource usage, and store them in the database;

[0040] Step 5-3: Establish accuracy, recall, F1 value, and average precision to evaluate recognition accuracy, use processing time to measure processing speed, and use memory and computing resource usage to evaluate resource consumption; based on data and indicators, adjust model parameters such as confidence.

[0041] The beneficial effects of the present invention are as follows:

[0042] Under the condition of meeting the actual accuracy requirements, compared with the method of realizing key part detection by relying on simple target detection and segmentation, the present invention can greatly reduce the amount of calculation and the labeling workload required for the preliminary tasks, reduce the algorithm's requirements for hardware and data volume, and thus reduce the cost during actual deployment. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 It is the system framework and process diagram of the present invention;

[0044] Figure 2 This is a diagram introducing the overall network architecture and functions of the present invention;

[0045] Figure 3 Examples of the results of the region perception and attention allocation process of the present invention, (a) coarse granularity, (b) standard granularity, and (c) high granularity;

[0046] Figure 4 Schematic diagram of training results according to an embodiment of the present invention;

[0047] Figure 5 These are examples of experimental results of embodiments of the present invention, including (a) rail and obstacle recognition, (b) ship, island, and hull recognition, and (c) tank and turret track recognition. DETAILED DESCRIPTION

[0048] The present invention will be further described below with reference to the accompanying drawings and examples.

[0049] The present invention aims to resolve the contradiction between computational efficiency and resource constraints in the identification of targets and their key parts in embedded systems, as well as the difficulty of maintaining high-precision identification in complex scenarios. By introducing regional perception and adaptive attention mechanisms, efficient and accurate recognition of target key parts is achieved, which is particularly suitable for resource-constrained environments and reduces system deployment costs.

[0050] This invention achieves accurate recognition of key parts of a target by introducing efficient region perception technology and adaptive attention mechanism. The specific objectives are as follows:

[0051] (1) Improve computational efficiency: By intelligently focusing on the target area rather than the entire image and directly utilizing the detection network feature information, the computational complexity is significantly reduced and the processing efficiency is improved. This is suitable for environments with limited computing resources, such as embedded systems.

[0052] (2) Enhanced robustness: Even in complex scenarios, such as lighting changes or background clutter, it can maintain high-precision recognition of key target parts and reduce the impact of external factors on the recognition effect.

[0053] (3) Optimize resource allocation: Use the adaptive attention module to perceive the key parts of the target and use feature information for regional perception rather than pixel-level analysis to avoid redundant processing of background information and reduce unnecessary computational overhead.

[0054] In summary, the goal of the present invention is to provide an efficient and accurate target and key part identification solution, which is particularly suitable for resource-constrained environments, so as to address the shortcomings of existing technologies in computing efficiency, real-time processing capabilities, sensitivity to external factors, and cost control.

[0055] The steps of the inventive method are as follows:

[0056] Step 1: Target detection and region preselection;

[0057] First, an object detection algorithm is used to perform global detection on the input image to locate all possible objects. This step outputs the bounding box coordinates of the object, which serves as the basis for subsequent region perception.

[0058] Objective: To ensure the accuracy of the initial positioning of the target and provide an effective regional scope for subsequent detailed analysis.

[0059] Technical highlights: Utilize the feature extraction capabilities of the deep learning model, combined with the anchor box mechanism and non-maximum suppression algorithm to obtain high-precision target bounding boxes.

[0060] Step 2: Region perception and feature extraction;

[0061] Based on the results of target detection, we intelligently focus on the area where each target is located, avoiding unnecessary calculations on the entire image. Using the feature map generated during the detection network inference, combined with the detection results in step 1, we perform further feature extraction and information fusion.

[0062] Goal: By directly using the feature maps generated by the inference of the detection network, the amount of computation can be greatly reduced and the processing speed can be increased while retaining enough information for the identification of key parts.

[0063] Technical highlights: A lightweight feature extraction module is used to directly obtain features from the detection network in step 1, reducing redundant calculations and optimizing resource utilization.

[0064] Step 3: Adaptive attention allocation;

[0065] Description: Based on the extracted feature information, an adaptive attention mechanism is used to dynamically adjust the attention distribution, so that the key part recognition model pays more attention to the key parts of the target rather than the background or secondary areas.

[0066] Objective: To ensure that computing resources are allocated preferentially to the areas with the most information value, and to improve the accuracy and efficiency of key area identification.

[0067] Technical highlights: Through a learnable weight matrix, attention weights are automatically assigned according to the statistical characteristics of the feature map to achieve focus on key areas.

[0068] Step 4: Identification and analysis of key parts;

[0069] Description: Use the features filtered by the attention mechanism to identify and analyze key areas.

[0070] Objective: To maintain or improve the accuracy and robustness of key part recognition while reducing computing resource consumption.

[0071] Technical highlights: Combining the predictive capabilities of deep learning models and the guiding role of attention mechanisms, feature extraction, precise positioning, and classification of key parts are performed.

[0072] Step 5: Result output and optimization feedback;

[0073] Description: Integrate the identified key parts information to form the final recognition result, and collect data during the execution process for continuous optimization and parameter adjustment of the model.

[0074] Objective: To provide accurate and timely key part identification results, and at the same time, gradually improve the performance and adaptability of the system through a feedback mechanism.

[0075] Technical points: Establish model evaluation indicators, monitor recognition accuracy, processing speed and resource consumption, and adjust algorithm parameters in a timely manner to meet the needs of different scenarios.

[0076] The specific measures and effects of each step are as follows:

[0077] Target detection and region preselection;

[0078] After acquiring the original or image-conditioned image from the image information source, a deep convolutional neural network (CNN) performs multi-layer convolution and pooling operations on the input image to extract feature maps at different levels. This automatically learns the characteristic patterns associated with the target in the image, effectively capturing the target's characteristic information. Based on prior knowledge of the target's shape and size, a variety of preset anchor boxes are used to increase the probability of detecting targets of different sizes and shapes, thereby improving the detection recall rate. The non-maximum suppression algorithm is used to reduce duplicate detections and improve detection accuracy. By suppressing bounding boxes with high overlap and low scores, the final output bounding box is ensured to be more representative and accurate. Based on these methods, accurate preliminary positioning of the target in the input image is achieved, providing an effective region range for subsequent detailed analysis, and improving the performance and accuracy of target detection and region preselection.

[0079] Region perception and feature extraction;

[0080] During the region perception stage, by focusing only on the target region, comprehensive image processing is avoided, significantly reducing computational effort. Because most background areas are often irrelevant to the identification and analysis of specific targets, focusing on key regions improves computational efficiency. Based on the target region bounding box determined by object detection and the feature maps generated during inference by the detection network, a specially designed lightweight feature extraction module processes the corresponding regions. These feature maps, learned and extracted through the network, contain semantic and spatial information, serving as the basis for subsequent feature extraction and analysis. Through weighted summation, concatenation, and the use of an attention mechanism, features at different levels are fused according to specific rules to obtain a more comprehensive and representative feature representation. By sharing feature information with the detection network and implementing region perception based on the detection results, computational effort is reduced and processing speed is increased during region perception and feature extraction, while retaining key features for subsequent identification of key target parts.

[0081] Adaptive attention allocation;

[0082] The extracted feature information, typically in the form of a multidimensional feature map, is analyzed and processed. By introducing a learnable weight matrix with the same dimensions as the feature map, the eigenvalue at each position is multiplied by the element at the corresponding position in the weight matrix to obtain a weighted eigenvalue. The weighted feature map is then normalized. The principle of the adaptive attention mechanism is to dynamically allocate computing resources based on the importance of the input features. By learning and identifying key patterns and information distribution in the feature map, the model can automatically focus more attention on the areas most valuable for key part identification. This mechanism can effectively capture local and global dependencies in the feature map, highlighting important features and suppressing the influence of irrelevant or minor features, thereby improving model performance and efficiency.

[0083] Assume that the input feature map is F∈R H×W×C , where H and W represent the height and width of the feature map respectively, and C represents the number of channels. The learnable weight matrix is ​​A∈R H×W×C .

[0084] The calculation process of attention allocation can be expressed as: F attended =F⊙A, where ⊙ represents an element-wise multiplication operation. Normalization is then performed. By continuously optimizing the weight matrix A, the model can learn the optimal attention allocation strategy during training, thus achieving adaptive attention allocation, focusing more computing resources on key areas, and improving the accuracy and efficiency of key area recognition.

[0085] Identification and analysis of key parts;

[0086] Using processed features, we obtain perception and identification of key parts. We use a combination of regression analysis and classification algorithms to locate and classify key parts. Regression analysis predicts the position and shape of key parts, while classification algorithms determine their categories. By leveraging the powerful representation and feature learning capabilities of deep learning models, combined with the guidance of attention mechanisms, we can accurately identify and locate key parts from complex input features.

[0087] Result output and optimization feedback;

[0088] When integrating and outputting results, organize key information into a unified data structure for clear presentation. Perform data validation and exception handling before output. Collect various data during execution, including input, intermediate and final results, system runtime, and resource usage, and efficiently store them in a database. Evaluate recognition accuracy using metrics such as accuracy, recall, F1 score, and mean average precision. Use processing time to measure processing speed, and memory and computing resource usage to assess resource consumption. Based on this data and metrics, adjust model parameters such as confidence levels.

[0089] Figure 2 Introducing the overall network architecture and functions

[0090] The method of the present invention can reduce the requirements for hardware and deployment costs, and can more efficiently obtain information on key parts of the target compared to simple target detection.

[0091] Example:

[0092] To verify the effectiveness of the algorithm in this paper, the results of feature information utilization and detection algorithm are shown below, as well as the actual implementation effect diagram in the final practical application process. According to actual needs, the algorithms in each link can be adjusted to enable it to have multiple capabilities such as target detection, key area detection, and parameter adaptation.

[0093] Illustration of the results of the regional perception and attention allocation process, for example Figure 3 As shown:

[0094] Figure 3 As a result diagram of the process of using different fine-grained regional perception and attention allocation for different scenarios in an untargeted scene, for example, coarse and fine-grained regional perception and attention allocation are used in the remote sensing field to focus on non-water areas, standard fine-grained regional perception and attention allocation are used in the aerial photography field to focus on the differences in characteristics of different regions, and high-grained regional perception and attention allocation are used in the road traffic field to ensure safety, focusing on suspected obstacle areas.

[0095] This paper uses mAP (mean Average Precision) as an evaluation metric for detection. mAP is the average of the APs of each category, which can reflect the average precision of the model; AP is the area under the PR curve, which is used to represent the average precision of the category. The PR curve is the Precision-Recall curve, where the calculation formulas for Precision and Recall are as follows:

[0096]

[0097] The above network is iteratively trained using the test dataset. The model results of each round of training are saved and the loss value is visualized. After the model training is completed, it is tested using the test set to obtain the mAP value of the model. The loss value, mAP, Precision, Recall and Precision-Recall curve corresponding to the training are as follows. Figure 4 shown.

[0098] The experimental results of the method of the present invention are exemplified as follows: Figure 5 , presenting the results for identifying railroad tracks and obstacles, ships, islands, and hulls, and tanks, turrets, and tracks. In actual deployment, the algorithm can be trained based on specific target datasets and fine-tuned to the scenario, fully leveraging the generalization capabilities of deep learning technology and reducing deployment costs.

[0099] Through the aforementioned steps, the method of the present invention achieves efficient and accurate identification of key target parts. From target detection to regional perception, attention allocation, and finally key part identification and analysis, each step aims to optimize the use of computing resources and improve recognition accuracy and robustness. It is particularly suitable for resource-constrained environments such as embedded systems. Through continuous data feedback and model optimization, the method of the present invention can continuously improve its performance in complex scenarios, providing strong technical support for fields such as intelligent monitoring, automated driving, and medical image analysis.

Claims

1. A method for identifying key areas using region perception, characterized in that: The steps include: Step 1: Target detection and region preselection; Use the target detection algorithm to perform global detection on the input image to locate all possible targets; Output the bounding box coordinates of the target; Step 1-1: After obtaining the original or image-conditioned image from the image information source, a deep convolutional neural network is used to perform multi-layer convolution and pooling operations on the input image to extract feature maps at different levels, automatically learning the characteristic patterns related to the target in the image, and thus capturing the target feature information; Step 1-2: Based on the prior knowledge of the target shape and size, by presetting a variety of different anchor boxes, the possibility of detecting targets of different sizes and shapes is increased, thereby improving the recall rate of detection; Step 1-3: Use the non-maximum suppression algorithm to suppress bounding boxes with high overlap and low scores to ensure the accuracy of the final output bounding box; Step 2: Region perception and feature extraction; Based on the results of target detection, focus on the area where each target is located; use the feature map generated during the detection network inference, combined with the detection results in step 1, to perform feature extraction and information fusion; Step 2-1: Based on the target area bounding box determined by target detection, combined with the feature map generated during the detection network inference, the corresponding area is processed using a lightweight feature extraction module; Step 2-2: Through weighted summation, concatenation, and the use of attention mechanisms, features at different levels are fused according to certain rules to obtain a more comprehensive and representative feature representation; Step 2-3: By sharing feature information with the detection network and achieving regional perception based on the detection results, key features are retained for subsequent target key part identification; Step 3: Adaptive attention allocation; Using an adaptive attention mechanism to dynamically adjust attention distribution, the key part recognition model focuses on the key parts of the target rather than the background or secondary areas; Step 3-1: Introduce a learnable weight matrix with the same dimension as the feature map. For the eigenvalue at each position, calculate its product with the element at the corresponding position in the weight matrix to obtain the weighted eigenvalue. Assume that the input feature map is F∈R H×W×C , where H and W represent the height and width of the feature map respectively, C represents the number of channels, and the learnable weight matrix is ​​A∈R H×W×C ; The calculation process of attention allocation is expressed as: F attended =F⊙A, where ⊙ represents an element-wise multiplication operation; Step 3-2: Normalize the weighted feature map; Step 4: Identification and analysis of key parts; Use the features filtered by the attention mechanism to identify and analyze key areas; Step 5: Result output and optimization feedback; The identified key parts information is integrated to form the final recognition result, and data from the execution process is collected for continuous optimization and parameter adjustment of the model.

2. The method for identifying key parts using region perception according to claim 1, characterized in that: The step 5 is specifically as follows: Step 5-1: Organize key part information into a unified data structure, and perform data verification and exception handling before output; Step 5-2: Collect various data during execution, including input, intermediate and final results, system running time and resource usage, and store them in the database; Step 5-3: Establish accuracy, recall, F1 value, and average precision to evaluate recognition accuracy, use processing time to measure processing speed, and use memory and computing resource usage to evaluate resource consumption; Based on data and indicators, adjust model parameters such as confidence.

Images