Ai mirror
An Ai mirror that looks at what you are wearing and reads it back to you.
Introduciton
Introduciton
Based on our interviews with Joe, we found a few areas where he struggled and we could be of assistance. One of these areas where his ability to pick out clothing. He struggles with knowing what is what and how the different pieces of clothing look and fit. There are many different ways of solving this problem, however one that the group came up with is an “Ai mirror”. The idea is to create what is essentially a mirror. A camera would take a picture of joe, then it would use AI to analyse what he is wearing and read it back to him. There are two main parts to this system. The first is the camera module. This would most likely be a wall-mounted, physical product that would house a camera and necessary electronics. Secondly, we have the Ai aspect of the product. This part of the product would needs to be connected to some form of existing large language model. We found that Chat Gpt 4 has recently gained the ability to analyse images(Beebom, 2023). This way we can automate the upload process and have Chat gpt describe the clothes in the image.
Based on our interviews with Joe, we found a few areas where he struggled and we could be of assistance. One of these areas where his ability to pick out clothing. He struggles with knowing what is what and how the different pieces of clothing look and fit. There are many different ways of solving this problem, however one that the group came up with is an “Ai mirror”. The idea is to create what is essentially a mirror. A camera would take a picture of joe, then it would use AI to analyse what he is wearing and read it back to him. There are two main parts to this system. The first is the camera module. This would most likely be a wall-mounted, physical product that would house a camera and necessary electronics. Secondly, we have the Ai aspect of the product. This part of the product would needs to be connected to some form of existing large language model. We found that Chat Gpt 4 has recently gained the ability to analyse images(Beebom, 2023). This way we can automate the upload process and have Chat gpt describe the clothes in the image.
Interaction
Interaction
The interaction between a visually impaired user and the physical product has have the necessary affordances. Firstly a way of activating the camera has to be developed. One way is ofcourse a button. This however means that there will have to be a delay and the users will have to know where to stand. Secondly there could be noise activation. A clap or some other form noise could be effective. This interaction however needs to be user tested.
The interaction between a visually impaired user and the physical product has have the necessary affordances. Firstly a way of activating the camera has to be developed. One way is ofcourse a button. This however means that there will have to be a delay and the users will have to know where to stand. Secondly there could be noise activation. A clap or some other form noise could be effective. This interaction however needs to be user tested.
Challenges
Challenges
This system presents a few challenges. Firstly there is the matter of privacy. Any active camera or listening device will always come with rightful concerns over privacy. Secondly there is the technical challenge. Automating the entire process and communication with chat gpt presents a significant challenge. Lastly there is as previously mentioned the challenge of interaction design.
This system presents a few challenges. Firstly there is the matter of privacy. Any active camera or listening device will always come with rightful concerns over privacy. Secondly there is the technical challenge. Automating the entire process and communication with chat gpt presents a significant challenge. Lastly there is as previously mentioned the challenge of interaction design.
Dsu group 3
Q2, 2024
