Jan 9-16 Journal
After our visit to CalTech last Thursday, we narrowed down our semester 2 project ideas to the four that will be elaborated upon below. Before I do that, here's what I have been doing since last Thursday. First, I have been looking into current e-cigarette vapor sensors and how thermal imaging recognition works. I have also discussed the different project ideas with group mates and am working with Robert to focus on this more ambitious project. I am looking forward to exploring how to gather a dataset of thermal images for algorithm training and how effective our sensor would be.
Idea 1: E-Cigarette Thermal Sensor Alert System
One major issue that afflicts the youth of today is the use of e-cigarettes, wiping away remarkable progress in fighting smoking. This rise not only creates another issue for health professionals but issues in a new wave of modernization in our response to high-tech nicotine. Due to privacy laws, most students use use e-cigarettes in school restrooms and lack of supervising personnel in schools hinder administration from tackling the issue. Current sensors for e-cigarette vapors exist, yet they rely on chemical based reactions activated by water vapor in e-cigarettes. Two major vulnerability in these current systems is that they first require exposure to the air, meaning increased possibilities of students tampering with/damaging equipment, and secondly are easier to avoid. When students position themselves or block the vapors, sensors may be less likely to detect e-cigarette use.
The solution we propose develops a better detection system for school administrators to identify use of e-cigarettes without a huge drain on resource. With a system that utilizes thermal sensors and image recognition through machine learning, a tamper-proof and more perceptive sensing system can be created. Thermal cameras only detect the heat signature of people, so privacy remains intact. Using thermal images, the system will identify instances of vaping by detecting the cloud that forms from e-cigarette use. The thermal camera will capture the cooler cloud as the temperature difference distorts normal body temperature. This anomaly will then result in a notification to administration for proper repercussions. To avoid errors in notifications, machine learning will be used for the system's artificial intelligence to identify e-cigarette clouds. A dataset of varying thermal images of e-cigarette vapors will train the algorithm, and this will result in improved reliability. The artificial intelligence will be aware of many different situations of vaping, so this will accommodate for student attempts to evade detection. It will also understand when certain temperature shifts do not match vaping clouds.
Idea 1: E-Cigarette Thermal Sensor Alert System
One major issue that afflicts the youth of today is the use of e-cigarettes, wiping away remarkable progress in fighting smoking. This rise not only creates another issue for health professionals but issues in a new wave of modernization in our response to high-tech nicotine. Due to privacy laws, most students use use e-cigarettes in school restrooms and lack of supervising personnel in schools hinder administration from tackling the issue. Current sensors for e-cigarette vapors exist, yet they rely on chemical based reactions activated by water vapor in e-cigarettes. Two major vulnerability in these current systems is that they first require exposure to the air, meaning increased possibilities of students tampering with/damaging equipment, and secondly are easier to avoid. When students position themselves or block the vapors, sensors may be less likely to detect e-cigarette use.
The solution we propose develops a better detection system for school administrators to identify use of e-cigarettes without a huge drain on resource. With a system that utilizes thermal sensors and image recognition through machine learning, a tamper-proof and more perceptive sensing system can be created. Thermal cameras only detect the heat signature of people, so privacy remains intact. Using thermal images, the system will identify instances of vaping by detecting the cloud that forms from e-cigarette use. The thermal camera will capture the cooler cloud as the temperature difference distorts normal body temperature. This anomaly will then result in a notification to administration for proper repercussions. To avoid errors in notifications, machine learning will be used for the system's artificial intelligence to identify e-cigarette clouds. A dataset of varying thermal images of e-cigarette vapors will train the algorithm, and this will result in improved reliability. The artificial intelligence will be aware of many different situations of vaping, so this will accommodate for student attempts to evade detection. It will also understand when certain temperature shifts do not match vaping clouds.
This idea utilizes haptic gloves to sense hand motions to hopefully translate ASL (American Sign Language) into captions. This will hopefully be used to teach people ASL more effectively while possibly being useful for deaf individuals to communicate. The idea is based on using haptic gloves which are used by virtual reality. These gloves have several sensors throughout them that track movements and we think that these gloves can sense the finger and hand motions of ASL. Machine learning would be applied to teach the glove system what motions translate into as not everyone will have the exact same movements.
Idea 3 + 4: Food Recommender System and Simple Game AI
Our third and fourth ideas are our two, more simpler ideas that will serve as a backup to the more ambitious first and second ideas. The food recommender system utilizes Yelp to create a dataset with matrices. We will then recommend several restaurants to people when they answer a quiz/survey that provides initial information for the AI. Will today found a program that automatically converts Yelp reviews into a spreadsheet, but with so many restaurants in our area, we plan to limit the program to a radius of 3 miles from the school. The simple game AI will be a program that is trained to play a game like tic-tac-toe or connect 4. These are more simpler games that will be easier to condense and train an algorithm.
Comments
Post a Comment