MSc Thesis Proposal Announcement by Puneet Jain " Enhancing Marijuana Intoxication Detection techniques by using deep learning-based architecture and image augmentation."

 

Date: Thursday 25 May 2023

Time:  09:30 am to 11:30 am

Location: Essex Hall, Room 122

 

Reminders: 1. Two-part attendance mandatory (sign-in sheet, QR Code) 2. Arrive 5-10 minutes prior to event starting - LATECOMERS WILL NOT BE ADMITTED. Note that due to demand, if the room has reached capacity, even if you are "early" admission is not guaranteed. 3. Please be respectful of the presenter by NOT knocking on the door for admittance once the door has been closed whether the presentation has begun or not (If the room is at capacity, overflow is not permitted (ie. sitting on floors) as this is a violation of the Fire Safety code). 4. Be respectful of the decision of the advisor/host of the event if you are not given admittance. The School of Computer Science has numerous events occurring soon. 

 

The increase in the number of motor vehicle and workplace accidents because of marijuana consumption has caused an economic impact on many countries. Δ-9-tetrahydrocannabinol (THC) is the primary psychoactive constituent of the Marijuana plant, and consumption of marijuana could result in red eyes. In this research, we have used bloodshot or red eyes after marijuana consumption as a potential indicator of marijuana use. The main problem with this research is the presence of a dataset, as there is no publicly available dataset. Although, this problem was tackled when the dataset was acquired by Raj in his thesis, Marijuana Intoxication Detection Using Convolutional Neural Network, by extracting images from Google searches and YouTube videos. Using this dataset, Raj achieved an accuracy of 82% by using a small CNN, MobileNet and traditional augmentation techniques such as rotation, random crop, flipping, and altering image contrast.

 

In existing research, work needs be done to increase the dataset using deep learning-based image augmentation, such as StyleGAN3. We will be using StyleGAN3 to increase the image dataset and improve the performance of the existing marijuana intoxication model by using state-of-the-art models like VGG-16, ResNet-50 and Inception-v3.

 

Keywords: Marijuana Intoxication detection, StyleGAN3, CNN, Image Augmentation

 

Internal Reader: Dr. Imran Ahmad
External Reader: Dr. Huapeng Wu
Advisor: Dr. Dan Wu