Week 6 Journal

After spending last week writing out our own ICA algorithm, the next step was to put our data we were collecting from the OpenMV into a format that we could use for the ICA algorithm. The ICA requires a matrix input, and the original code was tracking each of the channels in separate arrays. For this, I used the umatrix library I downloaded from GitHub, and found a way to append all three channels into one long array, and then rearrange the size into a 2D matrix, with 3 rows - representing the RGB channels.

Once that was done, the next portion of the algorithm was a DFT algorithm. Once the ICA algorithm was implemented, the RGB channels will be fully preprocessed. In order to extract the heart rate from these channels, the most efficient way we could think of was by taking a DFT. Unfortunately, OpenMV does not have an FFT function, so I looked into files that other people may have uploaded for this purpose. However, I could not get it to work, so I decided to code my own DFT function instead. Since OpenMV does have a math library (for real and complex numbers), I coded a DFT function, similar to the first lab in 113DA.

Next week, I will test the DFT function to make sure it is accurate. By then, Jen should have finished her helper functions and we will be able to test our ICA algorithm and debug accordingly.