Cross Talk
As camera angles change, joint angles can ‘disappear’ or one can be read as another. Anatomical or joint angles are measured in 3 dimensions around the three axes of each segment. For the knee, we typically first measure knee flexion about an axis that goes from side to side through the bottom of the femur (thigh bone). We then measure knee ab/adduction about an axis that runs from front to back, again through the bottom of the femur. (Note that rotation about this axis is typically very small; the knee joint is not designed to “flex sideways”.) Finally, we measure internal/external rotation about an axis that runs up/down the length of the lower leg bone (tibia).
When the projection angle is correct (meaning it equals the actual 3D joint angle) and the axes of both segments stay aligned (meaning there is only one kind of motion/rotation), then a projection angle will equal an actual joint angle. This happens pretty rarely in human movement. Let’s take a look at what happens to the measured “projection angle” when the camera viewpoint changes.
You've seen how important the position of the camera relative to the motion of the body can be and how a change in either camera position or direction of movement can affect the apparent amount of motion.
There is another example of this that we would like to show you where one kind of motion gets interpreted as another. In this case, thigh rotation will appear to affect knee flexion. In biomechanics, this is called crosstalk.
We're going to start with an overhead view to first show you how the thigh is moving during the downswing. As you can see in this golfer, the thigh starts in line with the foot and then rotates in substantially during the downswing. This movement, the combination of thigh adduction and internal rotation, can be 40 degrees or more just in the early downswing and substantially changes the relative angle between the leg and the camera.
So if we wanted to assess knee flexion, we might start by using this view, and it would look like the knee was flexed this amount. But as the golfer rotates his thigh, even though it is the same frame and we know that the knee angle is not changed, it makes it look like there is only this much knee flexion.
A golfer allows his trail leg to rotate out, externally rotate, during the backswing will look like he has less knee flexion than a golfer who has the same amount of knee flexion but keeps the knee pointing forward. Similarly, a golfer who rotates her thigh in during the early downswing, would look like she is either straightening her knee or bending her knee, not because she is, but because the orientation between the knee and the camera is changing.
We have shown there are a couple ways that a single camera view can provide an inaccurate view of the amount of motion. First, if the camera is not perpendicular to the movement (either rotation or translation) that movement will typically be underestimated. Second, some movements will change what is actually being measured. In the example above, when the knee is pointing at the camera, we get a reasonably accurately view of the amount the knee is abducted (or bent side to side) but as the thigh rotates inward, we will start seeing knee flexion, and we will now have a view that mixes abduction and flexion.
As you will see in the next module in the course, both of these problems can be avoided by using 3D systems because they are not limited to a single, fixed camera view. Instead 3D systems measure joint angles in all all three directions at all times, thereby giving a much more accurate description of what the knee, or any other joint, is actually doing during the swing.