Haptic Virtual Palpation of Volumetric CT: A Density-to-Force Rendering Method andIts Technical Characterisation

In a hospital radiology department, a doctor looks at a computer screen to see inside a patient's body. The images come from a CT scanner, a machine that uses X-rays to create a three-dimensional map of tissues. On this map, every tiny point has a number that tells the computer how much the tissue blocks X-rays. Dense bone blocks a lot and gets a high number; soft muscle blocks less and gets a lower number; air blocks almost nothing and gets a very low number. For decades, doctors have used their eyes to read these numbers and understand the body's structure. But when a doctor examines a patient in person, they do not just look; they touch. They press on the skin to feel if an organ is hard or soft, a crucial clue that a flat image cannot provide. The challenge for engineers has been to give a doctor the sense of touch while they are looking at a digital image, allowing them to "feel" the virtual bones and organs on a screen as if they were real.

A team of researchers at Jade University of Applied Sciences in Germany has built a system that attempts to solve this problem. They created a method that turns the density numbers from a CT scan directly into a feeling of resistance. When a user moves a special stylus, shaped like a pen and connected to a computer, across a virtual CT scan, the device pushes back against their hand. The harder the tissue is in the scan, the harder the device pushes back. This is not a simple trick of making everything feel the same; the system is designed to distinguish between different types of tissue, such as fat, fluid, soft organs, and bone, by changing the stiffness of the virtual surface in real time. The researchers wanted to know if this direct translation from image data to physical force actually works well enough to be useful, and if the computer can update the feeling fast enough to feel smooth and natural.

To test their idea, the team used a standard haptic device, a robotic arm that can push and pull a user's hand, connected to a computer running a simulation of abdominal CT scans. They did not use real patients or physical models of tissue. Instead, they programmed a virtual probe to move through five different public CT scans, tracing paths through the virtual body. As the probe moved, the computer calculated how much force to apply based on the density of the tissue at that exact spot. The system uses a specific rule to decide how stiff the tissue should feel. It treats the density numbers as a series of steps. For example, it assigns a specific stiffness to fat, a different one to fluid, and another to soft organs. Interestingly, the researchers programmed the system to make fluid feel softer than fat, even though fat is less dense than fluid in the scan data. This was a deliberate choice to match the physical reality that a probe sinks easily into free fluid but meets more resistance in fatty tissue, showing that the system prioritizes how a doctor expects to feel the body over a strict mathematical order of the numbers.

The results of the simulation showed that the system can successfully create a clear difference between soft and firm tissues. When the virtual probe moved through the softer bands, like fat or fluid, the force it felt was very light, measured in thousandths of a newton. When it moved into firmer areas, like soft organs or calcified tissue, the resistance increased significantly. The difference between the softest and firmest tissues was large enough to be clearly distinct in the data, with the firm tissues feeling roughly twenty to sixty-eight thousandths of a newton harder than the soft ones. This separation was consistent across all five different CT scans they tested. However, the researchers found a significant limitation when the probe touched the densest materials, such as bone. The device has a safety limit on how hard it can push to prevent it from shaking or vibrating uncontrollably. Because the bone in the scans was so dense, the force required to feel it properly hit this safety limit. As a result, the system could not fully show the difference between the very hard calcified tissue and the bone; both felt almost the same because the device was capped at its maximum push.

The study also looked at how fast the system could update the feeling. For a touch sensation to feel smooth and real, the computer needs to recalculate the force hundreds of times every second. The researchers found that their system updated the force between 213 and 278 times per second. While this is fast enough to be interactive, it is slower than the device's internal motor, which runs at 1,000 updates per second. This gap means that when the user touches a very stiff surface, the force feels a little bit stepped or jerky, which can cause a buzzing sensation. The researchers noted that if they had allowed the device to push harder, the difference between the dense tissues would have been much clearer, but they were forced to keep the force low to avoid that buzzing. They concluded that the method works well in a computer simulation to separate soft from firm tissues, but the current safety settings hide the differences between the hardest tissues.

Ultimately, the work demonstrates that it is possible to turn a CT scan into a tactile experience where a doctor can feel the difference between a soft organ and a hard bone. The system successfully translated the image data into a force that a human hand could perceive, creating a clear distinction between soft and firm areas. The main obstacle is not the way the image is converted into force, but rather the safety limits of the current hardware, which prevent the device from showing the full range of hardness found in the body. The researchers suggest that if they can measure the device's stability more precisely and allow for higher forces without the buzzing, the system could become a powerful tool for doctors to edit and verify 3D models of organs, adding the sense of touch to the visual information they already rely on.