AI-based Clinical Decision Support System
Design a practical artificial intelligence-guided clinical decision support system (AI-CDSS), by using several deep learning methods such as transfer learning on convolutional neural network (CNN) or other network structures to classify disease models.
Recently, several image-based diagnostic methods assisted by machine learning has been reported, and some of them show great performance. Those methodologies will be utilized to make a practical AI-CDSS, enabling to extract crucial features that medical doctors couldn’t recognize. Here, we are developing AI-based CDSS platform. Various medical images can serve as an input data set, and superior training network structure will be further investigated.
Human Robot Interaction (HRI)
HRI is a field of study dedicated to understanding, designing, and evaluating robotic systems for use by or with humans. Interaction is communication between robot and human
: Virtual reality (VR) technology has been highlighted recently in many industrial and biomedical fields. It has previously been applied to robotics micromanipulation, military skills in soldiers, phobia treatment, post-traumatic stress disorder, motor skill training on athletes, and surgical skill improvement. We research how to combine virtual reality technologies with human-robot interactions technologies. Since VR affords a more immersive experience than real environment, human robot interaction with VR can be used in training robotic surgery and remoting manipulator, which is safe and low-cost.
Visual Stimuli Modification in Virtual Reality
In the micro-surgical tasks, we assume that suppressing hand tremor is the first consideration to achieve successful surgery operation. We aim to assist a surgeon to have better self-awareness of hand tremor by modifying visual stimuli.
Here, it is presented two ways of visual stimuli modification using virtual reality: Modification 1: Tremor amplification and Modification 2: Object size magnification. By defining an orientation matching task, in which a subject is asked to match the orientation of virtual forceps to a given reference orientation. During the tasks, the subjects were instructed to put their elbow and lower arm on a table. For each trial, the subjects were requested to hold the forceps with the reference angle shown in the virtual reality screens (i.e., hologram image) for 20 seconds. The hologram reference image was in the same position and orientation throughout the tasks (yaw = 90º).
: Humans have some degree of hand tremor ranging from not noticeable in everyday tasks, to incapacitating tremor due to various neurodegenerative conditions.
: All surgeons strive to minimize tremor to successfully complete microsurgical procedures.
Optical Coherence Tomography (OCT)
: Optical coherence tomography has emerged as a dominant diagnostic imaging modality in clinical ophthalmology
: OCT can also be used as very precise distance sensor with high speed in biological applications, compared to other range-finding techniques such as ultrasound imaging or MRI.
Robot aided tool
: contains precise motor control to achieve surgical objectives and minimize surgical risks.
– C. Song, D. Y. Park, P. L. Gehlbach, S. J. Park, and J. U. Kang, “Fiber-optic OCT sensor guided “SMART” micro-forceps for microsurgery,”Biomed. Opt. Express 4(7), 1045–1050 (2013)
– C. Song, P. L. Gehlbach, and J. U. Kang, “Active tremor cancellation by a ‘smart’ handheld vitreoretinal microsurgical tool using sweptsource optical coherence tomography,” Opt. Express 20(21), 23414–23421 (2012)
Surgical Limitations of Micro-Surgery
– Physiological hand tremor about 6-12 Hz and 100 𝜇𝑚 motion
– Narrow eyesight
– Small size target
– Insufficient force feed back
Multi DOF manipulator which can transfer the end-effector to any direction and orientation is needed to compensate the physiological hand tremor and overcome surgical limitations
LINAPOD platform is a kind of 6 DOF parallel manipulator. This platform with very high stiffness employs drive modules and linear drives for linear motion. And the LINAPOD platform is used to develop robot-assisted handheld manipulator for microsurgery
Diffuse Speckle Contrast Analysis (DSCA)
In human health, stable blood circulation is very important for transporting nutrition, maintaining the temperature, and preserving cell-level metabolism. Diffuse speckle contrast analysis (DSCA) is in-vivo multi-channel blood flow measurement system with simple data analysis, flexibility, and fast measurement speed. This system estimates blood flow index (BFI) dependent on the correlation of speckle pattern caused by blood flow.
Lens based DSCA
In experimental setup, a laser source (Near Infrared) is connected with multi-mode fiber. A CCD camera mounted with biconvex lens (Magnitude=1)
Fiber-optic based DSCA
The laser fiber is coupled with a 1×4 coupler. Four optical fibers as detection channels are closely placed in front of the CCD chip in a CCD camera.
The system performance is evaluated using a phantom which is similar to optical properties of human skin. As the flow velocity from a syringe pump increases, the BFI shows increasingly larger value. Also, the DSCA system is applied to blood flow measurement on both human arm and chick embryos. In human arm experiment, the protocol has three periods including baseline, cuff-occlusion, and release, controlled by a blood pressure meter. At different periods, the BFI dramatically decreases and increases corresponding to the blood flow. In the chick embryo experiment, the BFI is measured during incubation period of chick embryos. As the embryo develops, the BFI gradually increases. In addition, the DSCA could determine the vital sign of the embryos. In conclusion, the DSCA system could secure in-vivo measurement of blood flow in deep tissues.
Mechanical properties measurement is useful for studying change or characteristics of biological system like cell or tissues. Elasticity is one of major biomechanical properties. For examples elastic properties of cancer cell are related to metastasis of cancer cell and diseased cell’s elasticity is changed.
Elastography is the method for measuring and imaging elasticity of sample by using various existing imaging modalities. OCE is one kind of elastography based on OCT system. While some force is applied to the surface of sample, OCT system measure and imaging surface displacement. From force and displacement information given by OCE system, Young’s modulus, signifying elasticity, is calculated.
Ultrasound is used in a lot of biomedical field because of its safety. In OCE system, ultrasonic force is used to give a force to the sample surface.
From the artery occlusion surgery, ischemic rat model would be used for stroke rehabilitation study.
Blood oxygenation/flow monitoring
By using modified Beer-Lambert law and scattering effect, non-invasive laser blood monitoring is available to distinguish stroke state.
Normally infarcted area indicates hypoxia symptoms than non-infarcted area.
Blood flow is also slower in infarcted area.
Designs optimal gait-training for better recovery of stroke
From the novel quantitative and non-invasive recovery level analyses,
we could modulate training intensity to get optimized training effect.
Precision Actuation and Sensing System
Determine number of turns of coil and thickness of leaf spring satisfying requirements of natural frequency and range which are given differently. Optimize voice coil motor system with other components such as driver, Dspace controller and capacitive sensor and check performance.
Microscale stent for rodent vasculature
Stent is frequently used to recover human vascular disease. The common stent for human has been developed. And, the development of stent for animal model is mostly focused on larger animals such as pig and rabbits. The rodent model has various advantages, involving economical approach, available to many genetic strains, and so on. We develop micro-scale stent for rodent with stable stent structure. To evaluate the structure, we observe the simulation depending on dilation, deflection, blood flow, as shown below figures. All simulation are tested using ANSYS software program. Then, the stents are practically validated on the rodents. Simultaneously, we apply assessment systems developed in our laboratory in order to monitor stent performance.