Abstract
Due to the mutual penetration and development of clinical medicine and informatics, medical image recognition can avoid the influence of subjective factors, and can diagnose the types of benign and malignant tumors in a timely and accurate manner, which is especially important for formulating effective treatment plans. This work mainly discusses fuzzy clustering and segmentation and SVM detection algorithms application in clinical medicine. The Internet of Things technology is a high-tech from the branch of the Internet, which plays a huge role in promoting the development and innovation of modern healthcare companies. The application of the Internet of Things technology has greatly changed the traditional medical model and effectively improved the relatively independent model in each unit system, thereby effectively promoting the scientific and informatization of modern intelligent medical care. Acute cerebral infarction is one of the most common clinical diseases, the clinical manifestations usually include tinnitus, headache, nausea, and vomiting. Acute cerebral infarction usually occurs suddenly and develops rapidly, which may eventually lead to hemiplegia, sensory disturbance, and language disturbance. This article analyzes the role of image recognition based on the medical Internet of Things in the clinical analysis of acute cerebral infarction and illustrates the clinical treatment methods through case studies. Simulation results prove that advanced IoT technology can more accurately track and monitor relevant patient information and can also play an important role in patient monitoring.
1. Introduction
With the development of electronic information and computer technology, medical imaging and image recognition technologies have developed rapidly, and have increasingly become powerful tools and key technologies for modern clinical analysis and medical diagnosis and treatment [1]. Although the imaging technology and equipment of modern medicine have made significant progress and improvements in recent years, the medical imaging film collected by modern medical equipment is still limited, and there are still many aspects such as low efficiency of the imaging engine of the equipment and imperfect technical materials. This is disappointing. For example, some medical imaging films have low spatial resolution, some medical imaging films contain a lot of noise and the signal-to-noise ratio is not good enough, or some medical imaging films must be combined with multipeak information for better clinical diagnosis. As an important branch of digital imaging technology, medical imaging technology combines technology and expertise in medical imaging, image graphics, and computer technology [2]. It can improve the visual quality of medical images and has been widely used in the medical film. The segmentation, selection, extraction, and recognition of medical images, as well as the completion of multipeak information fusion of medical images, have important practical significance in technology [3]. The research and development of medical imaging still faces many practical challenges [4]. For example, image recognition speed cannot meet the real-time requirements of medical clinics, and the accuracy of image segmentation has not been improved, because medical imaging technology has important practical significance for the research of medical clinical analysis and the diagnosis and treatment of clinical analysis [5]. In practice, patients with cerebral infarction have different manifestations, different symptoms, and different disease severity [6]. Some patients have mild symptoms, while others are severe. Acute cerebral infarction usually occurs suddenly and progresses rapidly, which may lead to hemiplegia, sensory disturbances, and language disturbances [7]. The disease usually peaks after 3–5 days. As the disease worsens, the damage caused by the disease becomes more serious, and the permanent damage to the nerves also becomes more serious [8]. Therefore, timely treatment is needed to effectively restore the blood supply and oxygen supply of the necrotic tissue and improve cerebral blood perfusion. Consistent with the current clinical drug situation, thrombolytic therapy with urokinase is effective and widely used, but some scientists still have doubts about drug treatment options [9]. Therefore, this article studies the clinical effect of urokinase thrombolysis in the treatment of acute cerebral infarction [10].
2. Related Work
The literature first introduces the background and importance of the subject research, then analyzes the current research status of brain image segmentation and recognition technology at home and abroad, and finally points out the main research content of this article [11]. The literature mainly provides general basic knowledge for brain image segmentation and recognition and introduces in detail the fuzzy sequence clustering c-means algorithm in the segmentation module, as well as the algorithm of support vector machine and recognition module related knowledge. At the same time, the advantages and disadvantages of the c-means fuzzy clustering algorithm are summarized, and the criteria for evaluating the effectiveness of the algorithm are given [12]. An improved FLICM algorithm based on neighborhood information is proposed in the literature. To accurately describe the influence of adjacent pixels on the central pixel in c-means fuzzy local clustering, it makes full use of the spatial information between local adjacent pixels and uses the gray information to improve the fuzzy control coefficient to change the value of the objective function [13]. And based on gradient descent, the iterative expressions of membership matrix and cluster centers are rederived. The simulation experiment proves that the algorithm improves the performance of the noise reduction algorithm under the premise of achieving the segmentation effect, and is suitable for the segmentation of medical images with high noise pollution [14]. The literature proposes an improved FLICM algorithm combined with SCoW. First, the image is presegmented using the SCoW method, and then it is split into many small superpixel blocks without destroying the edge details of the image. Then, the average feature of each small block is extracted as the input sample of the clustering, which reduces the computational complexity of the algorithm. Finally, the gray-scale correlation between the organizations is used to measure the similarity between the organizations and improve the segmentation accuracy [15]. Simulation experiments prove that the algorithm solves the problem of insufficient memory during operation and greatly improves the algorithm’s segmentation efficiency. The literature analyzes and summarizes the advantages and disadvantages of several typical algorithms with FCM extensions in fuzzy clustering algorithms [16]. To correct the disadvantages of the more traditional FLICM segmentation algorithm, refer to the ideas about improving other algorithms improved by FCM [17]. In addition, the literature also carried out in-depth research on two aspects of antinoise performance and segmentation efficiency, and then proposed an improved FLICM algorithm based on neighboring information, and combined with SCoW to propose an improved FLICM algorithm. Then, to more accurately distinguish brain tumors on brain images, the improved FLICM algorithm is used in combination with SCoW, and the brain tumors are determined according to the morphological, shape, and edge characteristics of the brain tumors after segmentation. The literature shows that the establishment of the medical Internet of Things platform solves the chaotic situation of the hospital’s original wireless network structure, the reuse of wireless network resources, the difficulties and increased costs caused by management and maintenance, and the frequent interference between wireless network systems [18]. The effect of hospital network coverage was improved.
3. Theoretical Basis of Image Recognition and Its Application in the Field of Brain Images
3.1. Overview of Related Algorithms for Brain Image Segmentation and Recognition
Assuming that X represents the gray-scale pixel value or characteristic pixel value of the image, the image is composed of area C, the center of the area cluster is the membership matrix, and the objective function is formula (1):
The objective cost function (1) satisfies the following conditions, such as formula (2):
The membership matrix and the expressions used for the iterative update of the cluster center can be obtained, such as formulas (3) and (4):
Considering the influence of neighboring pixels on the central pixel in the objective function, the objective function of FCMS algorithm is as formula (5):
The iterative update expressions of membership degree and cluster center are shown in formulas (6) and (7):
In order to further improve the performance of the FCMS algorithm, Chen et al. filtering technology based on FCMS is used to preprocess the pixel neighborhood information. The objective function is shown in formula (8):
Among them, such as formulas (9) and (10):
The objective function of EnFCM algorithm is as formula (11):
Among them, there is formula (12):
The expressions of the membership degree and iterative upgrade of the cluster center are as formulas (13) and (14):
The FGFCM algorithm measures the correlation between local pixels through the spatial and gray information between input pixels instead of α parameters, preprocesses the image, reduces noise while preserving image detail information, and effectively improves the performance of the image segmentation algorithm. Effect, the standard definition for measuring the similarity between local pixels, is as formula (15):
The gray-scale variance of adjacent pixels is defined as formula (16):
Then use formula (17) to filter the original image and obtain a new image defined as follows:
The FLICM algorithm does not need to manually select parameters, the clustering result is stable, and the target expression is shown in formula (18):
Among them, there is formula (19):
The expression of membership degree and the iterative update of cluster centers are shown in formulas (20) and (21).
According to the correlation between classes and between classes, this paper compares the division factor Vpc, which is defined as formulas (22) and (23):
For image segmentation, it is always hoped that pixels are located closer to the corresponding cluster center and far away from other cluster centers. Therefore, the Vxb scoring function based on the degree of association between classes is used to evaluate the segmentation performance of the algorithm. When the Vxb value is small, the clustering effect of the algorithm is better. Its definition is as formula (24):
Since most of the images collected in real life are mostly noise-contaminated images, this paper also uses the SA segmentation accuracy index as the index of the comparison algorithm, which is defined as the formula (25):
Support vector machine is a new pattern recognition technology developed based on years of statistical learning theory. It is usually used for machine learning problems, such as small samples, nonlinearity, high dimensionality, and feature fitting. The original main idea is to find the best classification area under the condition of linear separation. The basic description is shown in Figure 1. The green and purple dots in the figure represent the two training modes, and H is the interface that separates the two types.
If you continue to use the standard of the linear best classification area when the linearity is inseparable, it will inevitably lead to incorrect classification of a part of the data. At present, the quarantine zone in the strict sense no longer exists. Even in this case, we can still apply the maximum isolation band standard so that part of the data can fall into the isolation band or even the decision-making area of the other party, but the amount of this part of the data must be strictly controlled because the optimal classification corresponds to the area The schematic diagram is shown in Figure 2.
3.2. Improvement of Brain Image FLICM Segmentation Algorithm Based on Neighborhood Information
The so-called telemedicine is to monitor patients remotely through sensor technology and network technology to track all changes in the patient’s health and physiological laws. They are connected through the Internet. When the patient is in crisis, the patient’s family can quickly obtain the required information, so that it can be quickly transferred to remote monitoring. Experts and remote experts use the Internet to check the patient’s condition and quickly treat it. Get treatment, which saves time and helps patients get rid of life-threatening conditions. With the help of sensor technology, the corresponding sensor is connected to the patient, and the sensor is connected to the medical test equipment, so the doctor can remotely monitor, diagnose and treat the patient throughout the day. Nowadays, with the development of science and technology, it is possible to monitor the patient’s electrocardiogram, cardiopulmonary function, and breathing. At the same time, doctors can also determine the patient’s physiological characteristics and environmental characteristics. Comparison of segmentation results of different noisy images as shown in Figure 3.
(a)
(b)
(c)
Comparing the experimental results in Figure 3, this article also objectively confirmed the original intention of the original improved algorithm, and the experimental results are shown in Figure 4.
(a)
(b)
(c)
(d)
(e)
(f)
(g)
(h)
(i)
(j)
(k)
The clustering results and segmentation accuracy of the algorithm proposed in Figure 5 are significantly better than other improved FCM algorithms, which further proves that the algorithm proposed in this paper is robust and feasible.
3.3. Medical Image Segmentation Results
To better understand the advantages and disadvantages of each algorithm, the experiment in this article compares the distribution coefficient Vpc, the distribution entropy Vpe and the estimator Vxb according to the correlation between the three segmentation algorithms. The experimental results are listed in Table 1.
Table 2 also shows that the segmentation effect of the enhanced SCoW_FLICM algorithm in this paper is significantly better than that of the FLICM segmentation algorithm.
4. Clinical Analysis of Acute Cerebral Infarction under the Background of Intelligent Medical Treatment
4.1. Medical Internet of Things
At present, the development of smart medicine is rapid, but there are still some problems. For example, there are few talents with medical and Internet of Things technology, and the use of Internet of Things technology is full of certain security risks. Since doctors and patients are the main targets of smart medical services, the safety and privacy of both parties are very important, and the privacy of both parties must be respected. Despite the introduction of information management, the hospital information system still has some problems, such as the outpatient business process is still complicated, the information between departments is not smooth, and the information about patient diagnosis and treatment cannot be accessed and input. Moreover, detailed information about the circulation of purchased drugs cannot be obtained in time, and these problems severely limit the management of the hospital information system.
The main function of the electronic case is to record the identity of the patient, etc. The electronic medical record can collect information about the patient, process the collected information, and quickly and accurately understand the patient’s diagnosis in the hospital when necessary, and save it in the personal electronic medical record in the record. The doctor can fully understand the patient’s entire medical history, so that the existing problems can be solved effectively and accurately. Now all that is left is to use the Internet of Things to attach the corresponding RFID tag to the drug. Using this tag, you can fully understand all kinds of information about the drug, and it also has the corresponding anticounterfeiting function, so employees can quickly and accurately verify the drug. Drugs have greatly reduced the occurrence of health accidents.
The continuous updating and improvement of information technology has laid a solid foundation for smart medical care. The hospital information system can contribute to the normal operation of the hospital in the new century, and it is also an important technology to support the operation of the hospital. The main purpose of the hospital information system is the patient. The main purpose is to understand basic patient information, propose treatment plans and appropriately consider medical expenses, and integrate the healthcare, nursing and medical technology departments into the management information system, thus making us online Professionals-provide advice and telemedicine records.
4.2. Pathogenesis and Image Diagnosis Analysis in Patients with Acute Cerebral Infarction
In patients with acute cerebral infarction, within 6 hours after the onset, when the cerebral blood flow is reduced, the brain tissue remains ischemic and hypoxic. Therefore, in the hyperacute phase, it is difficult to accurately distinguish the infarcted brain tissue from the normal brain tissue, and the grayscale difference is the smallest. In patients with early cerebral infarction, changes in tissue function occur earlier than changes in morphology. DWI shows that the diffusion of water molecules is restricted, and the signal in the affected area is higher. Because CT is extremely sensitive to cerebral hemorrhage and calcification in patients with acute cerebral infarction, the main advantage of CT is that it can eliminate cerebral and thyroid hemorrhage.
Brain damage caused by decreased parathyroid function is of great significance for the further development of thrombolytic therapy with tissue plasminogen activator. However, for patients with acute and subacute diseases, the examination period starts from 6 to 24 hours, starting from 1 to 14 days, it gradually changes from the cytotoxic edema stage to the angioedema stage. At this time, the intracellular and extracellular water content increased significantly. CT and MRI images may show obvious abnormalities, including possible display of significant CT reduction, early shadow low density, unclear cortical and white matter boundaries, lenticular lens and pancreatic islet features, ventricles, and other changes, which were later visualized as a middle cerebral artery or high-density basement Signs of arteries. Among them, the early shadow of low density is a typical sign of cerebral infarction, and it is more common in the stage of cytotoxic brain edema. At this time, the boundary between the cortex and the white matter is fuzzy and uniform, the shadow is low-density, the boundary is fuzzy, and convolute, the groove of the ventricle and the change of the middle cerebral artery or the high-density area of the basilar artery occur relatively late. The first is caused by the expansion of brain tissue swelling, which only occurs when the brain sulcus, subarachnoid space, or even ventricular malformations and midline displacement are compressed. However, on CT, it is still difficult to detect brain stem and cerebellar infarctions in the posterior fossa. At that time, the number of abnormalities detected by CT was still affected by factors such as imaging time, so 9.17% and 4.78% of the cases were still detected, respectively abnormal. Although studies have shown that DWI still shows limited diffusion and high signal. The results showed that during the 3d inspection period, the injury was related to angioedema, collateral vessel formation, and a large number of swollen cells. At this time, the DWI signal is relatively attenuated, and this study also shows that MRI is in a worsening stage. The incidence of first cerebral infarction was 97.92% and 100.00%, which were still higher than CT. 2.08% of patients underwent MRI examination within 6–24 hours after the onset of the disease, which showed that the lesion was in the brainstem. Therefore, patients with acute cerebral infarction need CT and MRI examinations, which can be used for clinical diagnosis and treatment of cerebral infarction.
4.3. Clinical Analysis of Urokinase Thrombolysis in the Treatment of Acute Cerebral Infarction
During the study period from February 2018 to January 2020, 120 patients with acute cerebral infarction were included in the database as experimental samples. The control group consisted of 36 men and 24 women with an average age of (66.8 ± 5.90) years old. The patients in the experimental group consisted of 29 men and 31 women (65.8 ± 6.10) years old from 53 to 75 years old. The experimental group adopted medical Internet of Things equipment to conduct more timely treatment interventions for patients, and update information in real time. Study patients, patients with poor physical tolerance who cannot be examined, pregnant women, and patients receiving appropriate treatment for three months. Enter the date of admission of the patient into the database, and the test result shows ; the difference is not significant and can be compared.
The experimental results showed that the experimental group had shorter hospital stays in hospitals and intensive care units, and the difference was statistically significant , see Table 3.
The experimental results show that after the treatment, the total effective treatment rate of the experimental group of patients is 85.00% (51/60), see Table 4.
The experimental results showed that after the treatment, the adverse reaction rate in the experimental group was 6.67% (4/60), while that in the control group was 26.67% (16/60) after treatment, see Table 5.
The older the patient, the greater the likelihood of an increase in the incidence of this disease. The disease has a significant impact on the specific functions of the patient and, in severe cases, the normal cognitive function of the patient. As far as the current situation is concerned, choosing a neurology treatment plan is the main theme of modern clinical practice when clinical treatment is started, and the use of urokinase to treat patients has also become a new focus of clinical research in recent years.
Urokinase is a direct plasmin activator, which can effectively remove a hematoma, dissolve embolism, and has a good therapeutic effect on acute cerebral infarction (onset within 6 hours), and the drug is relatively safe. When treating patients, choosing urokinase for treatment can ensure that the patient’s infarction returns to normal and continue to obtain therapeutic effects. However, some studies have shown that urokinase can aggravate blood vessel damage and damage patients. When the patient receives urokinase treatment, the medication must be adjusted accordingly according to the patient’s condition so that the patient can obtain adequate medical support during the treatment process. Therefore, clinical conservative treatment or surgical treatment drugs can be appropriately selected according to the situation of patients. In order to effectively control the mortality rate in patients to avoid recurrence or complications conditions to choose. In this study, patients with acute cerebral infarction received Xueshuantong therapy and urokinase thrombolytic therapy for a comparative study.
Therefore, in the treatment of patients with acute cerebral infarction (onset within 6 hours), intravenous thrombolytic urokinase therapy may provide the best therapeutic effect, and in the case of actual use of urokinase, it should be taken as soon as possible to improve The overall condition of the patient. After the completion of the treatment, the effective treatment process for the patient and the reduction of the patient’s recurrence rate.
5. Conclusion
Since the future development of the network is usually related to the networking and informatization of human life in the environment, the status of the hospital information structure related to the Internet of Things technology has become important. Advanced Internet of Things technology can not only be applied to related advanced equipment to track and monitor relevant patient information more accurately, but also can play an important role in patient monitoring. At the same time, it greatly reduces the possibility of health accidents and ensures the effective improvement of medical performance. Acute cerebral infarction is a type of cerebral infarction that lasts for 14 days. At this stage, cerebral ischemia and hypoxic injury are reversible, and the restoration of cerebral blood perfusion can save the traumatic brain injury to the greatest extent. How to detect a heart attack in time, effectively evaluate its size, location, etc., and respond as soon as possible within the effective thrombolytic time window to reduce the infarct size, restore ischemic penumbra infarction, and prevent long-term ischemia, so that The main problem we want to study. The effect on the function of the nerve nucleus is particularly important for restoring the normal function of nerve conduction. In this study, patients with acute cerebral infarction received Xueshuantong therapy and urokinase thrombolytic therapy for a comparative study. Therefore, in the treatment of patients with acute cerebral infarction (onset within 6 hours), intravenous thrombolytic urokinase therapy may provide the best therapeutic effect, and in the case of actual use of urokinase, it should be taken as soon as possible to improve The overall condition of the patient. After the completion of the treatment, the effective treatment process for the patient and the reduction of the patient’s recurrence rate.
Data Availability
The data used to support the findings of this study are available from the corresponding author upon request.
Conflicts of Interest
The authors declare that they have no conflicts of interest.
Acknowledgments
This paper was supported by Research project funded by Health and Health Committee of Guangxi Zhuang Autonomous Region: Effect of urokinase intravenous thrombolysis on serum inflammatory factors and nerve function in patients with acute cerebral infarction (Z20190207).