FYP 1-2

e-portfolio evaluation entry 2

This is my FYP progress

After some literature review research, there are a lot of possible ways in creating the system. First, a hand gestures recognition offer many approach includes vision-based and non-vision based. Vision-based example is using camera while non-vision is using glove with sensors and displacement sensors. Second, the deep learning as recognizer, CNN architecture is the most popular and appealing as the convolution layer in the architecture will help the training and testing phase to yield a high accuracy. Lastly, smart home; there are not many research paper that focusing on both recognition and automation. But there some demonstration has been done using Arduino and Raspberry Pi.
Thus, based on the literature review made. CNN architecture will be used a the classifier of a static hand gesture that will be automate by controlling the Arduino components.
The flow of work is 1) datasets collection of hand gestures includes the preprocessing steps 2) training and testing the CNN architecture 3) design and test smart home part 4) implement both that include control units
in order to do the research the knowledge of deep learning is needed includes programming language and input device and also the knowledge of Arduino
The hardware generally Arduino such as Bluetooth model, switch relay and Arduino socket. Software need are Python and OpenCV

Updated on 11 February 2022, 12:48 PM; 143 page visits from 17 December 2021 to 11 October 2025