Researches Use IoT for Cancer Diagnosis

Researchers are using the Internet of Things to diagnose cancer
Researchers at the Korea Institute of Science and Technology (KIST) have developed a new technology for cancer diagnosis; a simple yet accurate method that uses tactile neuron devices combined with AI technology.

Usually, the non-invasive diagnostic method is ultrasound elastography; however, the interpretation of the results may differ. The new method identifies and measures the severity and spread of a tumor, enabling accurate cancer diagnosis.

The KIST team developed this alternative method to improve accuracy and speed up prognosis. For their experiments, the team combined tactile neural devices with artificial neural network learning methods that apply pressure to a potentially cancerous site with a strong force. which produces electrical spikes that increase or decrease depending on the hardness of the object encountered. The method falls under the category of “neuromorphic technology”, a data processing technology that is increasingly popular due to its compatibility with AI, IoT and autonomous technologies. It tries to mimic the way the human brain processes large amounts of information with less energy, with neurons receiving external stimuli through sensory receptors that then translate the signals into electrical spikes.

Used to diagnose the disease, the team used elastography images of malignant and benign breast tumors combined with a learning method for neural networks. The pixels of the color-coded ultrasound elastography image relate to the hardness of the object found and are converted to frequency values ​​for AI training.

After this process, the team reported a breast tumor diagnosis accuracy of 95.8% and said that the developed artificial tactile neuron technology is able to “identify and learn mechanical properties with a simple structure and method.”

The team also expects the device to be used in low-power, high-precision disease diagnosis and applications such as robotic surgery, where a surgical site must be quickly identified with little or no human interaction.