FYP 1-3

by MUHAMMAD FAIZ BIN MOHAMAD SUBZAN

E-portfolio entry progress evaluation 3

Litereture Review

This is the literature review progress until 6th January

File(s) to download

Text

  • In summary, MNIST dataset which is a handwritten digits is used to simulate the hand gesture that to be used in the research. In other word, to train the architecture a datasets is used. The CNN architecture consist of 3 layers of convolutional and max pooling layers followed with 3 fully-connected layers. The validation accuracy is 98%.
  • but no smart home implementation is made during preliminary.
  • The CNN can provide the recognizer for the hand gestures later in the research and the used to control the home automation.
Updated on 11 February 2022, 12:55 PM; 189 page visits from 9 January 2022 to 11 October 2025