Research Article
LSTM-Based Neural Network to Recognize Human Activities Using Deep Learning Techniques
Algorithm 1
 Removing duplicate image frames
| | (i) | Prerequisites: Generate image frames from input images and store them for further process. |  | (ii) | Step 1. Iterate through all the images of a directory |  | (iii) | Step 2. Generate the file hash using MD5 method for each image by dividing n-bit state into  bit words and label it. |  | (iv) | Step 3. Perform four stage n-bit message block process using nonlinear function F and store it in a list. |  | (v) | Step 4. Verify the image hash value in hash keys list. |  | (vi) | Step 5. If exists, mark the image as duplicate. |  | (vii) | Step 6. Remove the image from the dataset. |  | (viii) | Step 7. Repeat the process from step 2 to step 6 for other dataset images | 
 |