In my past blogs, I mentioned that I have been working on a project that recognizes footprint image captured by Tactonic pressure sensor, and obtains the Arch Index by calculating the ratio of the middle third area to the whole footprint area. Arch Index is a good way to measure the height of the foot arch; it is widely adopted in Physical Therapy since the height of the arch is a significant factor in balance stabilization.
The traditional way to obtain arch index is Harris Mat which stacks three layers: rubber layer on the top, paper layer in the middle and ink layer on the bottom. Testees stepping on Harris Mat can print their foot pressure image on the paper layer. Even if the testees immediately lift up their feet, the paper layer would record the augmentation image that contains all subtle movements in the minimal amount of time. It is just like fingerprinting; the scanner records the augmentation image along with your finger rolling procedure.
However, the current version of my arch index system detects arch index in real-time. It runs arch index detection on every frame, and the frame only represents the pressure information at the very moment, it is independent with its previous and next frame so it is not augmentation image like the Harris Mat can get. This explains that the result of our arch index results were slightly different than the ground true data we got from Harris Mat. I am going to improve my arch index program so that it can run arch index detection on augmentation images. I will update again when I get exciting progress on it.