LSTM-Based Neural Network to Recognize Human Activities Using Deep Learning Techniques

Algorithm 3

Feature extraction using mean pixel values

(i)	Step 1. Load the RGB image with a dimension width x height x c where “c” is the number of color channels
(ii)	Step 2. Generate the number of features as a product of width, height, and c.
(iii)	Step 3. Generate a new matrix by calculating the mean value of all the pixels
(iv)	Step 4. Create a new image with same width and height by initializing all the values to 0.
(v)	Step 5. Generate a 1D array by appending all the pixel values one after another.