• Alex Savin
• January 15, 2022
• Share Now

Computer Vision Healthcare and Medical Applications

Fifty years ago, there were 5,000 diseases, today there are 30,000. By the end of this century, it is estimated to be 60,000. Medical progress means that the number of laboratory tests in a year and medical knowledge will double within five years. Data processing devices are going to be one of the most important furnishings in a doctor’s office. Do not be scared, hospitals without doctors will appear only in 2035. Let’s dig into the world trends in computer vision healthcare. 

Healthcare Industry Trends Driven by Computer Vision

Today, the healthcare industry has formed a number of trends that affect not only the work of large corporations, insurance companies, and clinics but also the lives of each of us. One of the key trends in medicine is the constant increase in the cost of treating patients. Investment companies are interested in reducing the cost of treatment and services. AI and machine learning are absolutely essential for this. How CV will do this?

First, the introduction of an individual approach to treatment. This is an opportunity to improve the quality of treatment using several methods:

• monitor the patient’s condition, collect data about him;
• make a remote examination using devices that convey the patient’s condition;
• the ability to create an individualized treatment plan for each patient;
• early diagnosis.

The second is the use of support staff to advise patients. This saves doctors time. Nurses are attracted for the initial assessment of the patient’s condition, the detection of anomalies in the diagnostic results. Moreover, the elaboration of schemes and algorithms for treatment, the creation of questionnaires allows helping non-medical staff to determine the patient’s condition and decide whether he needs a doctor’s advice or not. The same algorithms allow the implementation of bots to process initial requests from patients.

Opportunities and The Use of Computer Vision in Healthcare 

Assisted systems

According to Anand Rao’s “A Strategist’s Guide to Artificial Intelligence,” assisted systems will be commercially available and used by 2020. Image classification systems help the doctor to conduct high-quality diagnostics with minimal time. Now the classification of medical images and the description of the pictures are done by radiologists, ultrasound specialists, etc. The analysis of the images can already be carried out using artificial intelligence automatically.

The patients at risk are determined with the help of artificial intelligence. The doctor pays attention to these patients first. Thus, time is significantly saved and the possibility of a mistake of doctors is minimized. For example, there are programs for detecting melanoma.

About 422 million people from different countries suffer from various types of diabetes. Every third diabetic sooner or later receives diabetic retinopathy, due to which a person can completely lose his sight if time does not begin treatment. IBM has decided to use a combination of technologies, including deep learning, ultra-precise neural networks and visual analytics, which together help diagnose retinopathy with 86% accuracy. In this case, the technology involved a base of 35,000 images of EyePAC. During the “refinement” of this technology, the company’s specialists identified the main markers by which retinopathy can be identified. First of all, it is damage to the blood vessels of the retina.

The duration of the screening procedure is only 20 seconds. During this time, the diagnostic system can determine retinopathy with a high degree of accuracy. As one of the tools used in the procedure for screening eyes is an ordinary mobile phone. Developers believe that the new technology should be used as an auxiliary tool that complements the capabilities of doctors.

Digital Diagnostics 

Now there is no need to visit the clinic for blood testing, weight control, etc. Now, even a smartphone with a health application can perform your body tests like blood pressure, pulse, insulin amount and so on. This automation in healthcare also made it possible to automatically send patients prompts about the correct time for their scheduled check-ups. The reduction of visits to the medical facility and the increase in the number of moments of remote health monitoring has created such an area as telemedicine. Many patients agreed that the home is the best place for treatment, while patients are in a “normal everyday environment.” 

The market of “smart” medical devices every year becomes more diverse in range and more in volume. This allows medical professionals to evaluate, diagnose and treat patients, using conventional technologies, such as telephones and computers, to provide clinical services to patients at a great distance. In particular, telemedicine is a huge success in the field of psychiatric care. Patients who need emotional support can call the therapist with one click.

In developing new devices, manufacturers are trying to improve them by adding additional functions, which leads to the production of multifunctional devices. So we can say that it is becoming increasingly difficult to classify such devices into a particular category. With the help of such devices, the user has the ability to monitor his physical condition, and if necessary – send the data to the doctor.

Internet of Medical Things (IoMT)

The Internet of Medical Things is a network of connected medical devices and solutions that collect data and transfer it to the IT system for further analysis. These are sensors, wearable devices, mobile applications, smart beds, tablets, smart measuring devices.

The number of connected medical devices is expected to increase from 10 billion to 50 billion over the next decade. Cisco estimates that by 2021, the total amount of data created by any IoT device will reach 847 zettabytes (ZB) per year. At some point, IoT will become the largest data source on Earth. Soon humankind-oriented data, such as medical history, drug allergies, laboratory test results, personal statistics, among many other things, will be digitized as part of electronic health objects. Doctors will be able to interpret and use a wealth of big data from connected systems to make informed decisions about patient care, as well as to understand and predict current and future health trends. The key to all this is Machine Learning.

An autonomous “nurse” is an example of a medical AI-enabled IoT application. She will be able to answer patient questions because she is connected via the Internet to a wide range of data from previous health records. By integrating facial recognition, a robotic “nurse” will be able to recognize the patient’s mood and adapt his behavior and reaction accordingly. She will also be able to remind her patients to take medicine and also remind them about prescribing a doctor.

Now imagine if the hospital “recruits” nursing robots that can reason, make choices, learn, communicate, move, and connect to the hospital’s network and are connected to each other, they can help the nurses with tasks like drug administration, records, and communication with doctors, patient education and disease management. This will be a good solution to many medical and organizational problems, especially in cases where nurses sometimes have to handle more than they can.

Epidemiology and preventive medicine

The health authorities would like to be better informed about the occurrence of certain diseases – be it a flu epidemic or a certain sexually transmitted disease. Computer vision in medical data inventories according to regional aspects bring considerable progress here. Humanity regularly encounters epidemics. Many have heard about Ebola, malaria, and other outbreaks of diseases that are transmitted both by insects and through the water. 

An artificial intelligence system that allows you to control and predict epidemics is in the status of clinical research, but it has already been used and is operating in Africa. Information is collected using drones. They actually catch mosquitoes, analyze their DNA and give a prediction: where and when the next epidemic will occur – after which the risk area is treated. Such systems help prevent uncontrolled outbreaks of epidemics.

Nuclear medicine

Nuclear medicine is a section of clinical medicine that deals with the use of radionuclide pharmaceuticals in diagnosis and treatment. Sometimes methods of remote radiation therapy are also referred to as nuclear medicine. In diagnostics, it mainly uses single-photon emission computed tomography (SPECT, capture gamma radiation) and positron emission tomography (PET scanners), and radioiodine therapy prevails in the treatment.

Nuclear medicine began with scintigraphy. The essence of this study is in the intravenous administration of a radiopharmaceutical that carries the radioactive isotope. Each type of drug accumulates in a strictly defined organ. If for scintigraphy for each organ they use their isotope, the technology named Positron Emission Tomography needs only one element – 2-deoxy-2-fluoro-D-glucose (FDG). It tends to accumulate in malignant cells. 

The main task of PET is to detect a tumor or its metastases at a time when other methods of research do not “see” them. However, in itself, PET technology gives a rather blurry visual result, according to which it is difficult to establish the exact location of the accumulation. Therefore, PET/CT (Computerized Tomography) is most often performed today, when computed tomography is performed along with positron emission, and then the PET result is superimposed on precise sections obtained by CT.

In 90% of cases, PET/CT is used to detect the recurrence of a malignant tumor or the appearance of its metastases in time. This allows you to start re-treatment on time – perform surgery, radiation, or chemotherapy, which gives a chance to finally defeat the disease. Doctors also use PET/CT in controversial cases when other methods do not allow to distinguish a benign tumor from a malignant one. 

There are situations when the analysis for tumor markers shows the presence of malignant processes, but even a detailed examination does not take specialists to its mark. In this case, PET/CT is indispensable. It is necessary to look for a neoplasm in the body where the drug accumulates.

Laparoscopic Surgery

In modern laparoscopic operations, the image on the endoscope is complemented by the image obtained during intraoperative angiography. This allows the surgeon to know exactly where the tumor is inside the organ, and thus minimize the loss of healthy tissue of the patient’s body during surgery to remove the tumor.  Traditional laparoscopic surgery without the use of medical imaging is not able to effectively carry out operations in which you need to see the internal structure of the organs. 

For example, if cancer in the kidney, liver, or pancreas is located inside the organ, and not on the surface, the surgeon is unable to see the tumor through the holes in the abdomen. Therefore, in recent years, laparoscopic surgery with the use of medical imaging, carried out in hybrid operating rooms, has become widespread. Computer vision healthcare methods enable a good quality of medical imaging. The ability to take images directly in the operating room and the ability to accurately direct surgical instruments during surgery contribute to the spread of this approach.

A new stage in the development of laparoscopic surgery was the use of specialized robots, one of the most famous of which is daVinci. This robot is equipped with micro tools, much smaller than standard laparoscopic instruments, as well as a miniature video camera that reproduces a color, three-dimensional image of the operation in real-time. The movements of the surgeon are transferred by the robot to the smooth movements of microtools, capable of moving in all directions. With their help, the operation is performed much more precisely, keeping intact the most delicate plexuses of the nerves and blood vessels.

Concluding Thoughts on Computer Vision Healthcare

The development of modern medicine cannot be imagined without the introduction of IT technologies. Digital transformation is gradually capturing all industries in the world, helping to solve problems that stand in the way of their development. Current trends in the healthcare sector have led to an increase in demand and, consequently, investment in the development of computer vision healthcare solutions.

Further development of artificial intelligence will lead to the use of computer vision healthcare augmented artificial intelligence systems. These systems open up new possibilities for us. For example, to categorize MRI images at high speed without human intervention. Also, create personalized medicine and effective treatment based on the patient’s specific data – analyzes and reactions to chemicals. According to forecasts, such a service will be available for mass use by 2030.

By 2035 we expect the emergence of hospitals without doctors. This is an example of autonomous artificial intelligence when the system itself makes decisions. Doctors of course will still be needed, but for some simple cases, the above-described CV technologies will be available.