Oscar Wang
UX + Creative Technology + Accessible
Hi , I am Oscar Wang,
a user experience designer, specializing in using creative technology along with physical devices to find design opportunities for underserved population.
© 2025 Oscar Wang
Type
AI, Programming, Computer Vision,
Machine Learning, UI Design
Date
2024.02
Participant
Monica Zhao, Oscar Wang
Alris
Peoject Introduction
In daily life, many people often suffer from the inability to take good photos. I hope to comprehensively solve this problem by applying AI, machine learning and computer vision.
What is Alris?
AIris uses advanced image recognition technology to analyze the surrounding environment through the camera, offering intelligent suggestions for the most visually appealing compositions, optimal angles, and dynamic poses.
It provides real-time, actionable feedback, ensuring each shot is captured with the highest quality through continuous optimization.
Technology Utilize
ChatGPT
We use chatpgt to give suggestions on the modification of poses and other problems in pictures, and present them to users in the form of text and summary.
Google Teachable Machine
We trained a model using the google teachable machine to analyze the strength and weakness of light sources in a video or image in real time.
Ml5: PoseNet & NeuralNetwork Database
We used posenet's skeleton analysis function to track the user's shape and movement, and built a database to store different photo poses.
Working Protoype Video
Project Poster
Prototype Flow
Before Taking a Photo
Before taking photos, Alris will use the model summarized by machine learning to analyze factors such as the Angle of the light source. The upper right corner will directly display the situation of the light source, and the lower right corner will give corresponding suggestions.
Taking a Photo
After taking a photo, the AI will analyze whether the photo is suitable according to the trained model, and directly display the photo is good or bad in the upper left corner, and display detailed analysis of the reason below. If the photo is bad, a pose suggestion button appears at the bottom of the screen.
Pose Suggestion
In pose suggestion, the AI finds appropriate actions in the database and recommends them to the user based on the photo analysis results, and scans the user with computer vision. If the user makes this action correctly, it will automatically take a photo of the user and save it after a few seconds.
Type
AI, Programming, Computer Vision,
Machine Learning, UI Design
Date
2024.02
Participant
Monica Zhao, Oscar Wang
Alris
Peoject Introduction
In daily life, many people often suffer from the inability to take good photos. I hope to comprehensively solve this problem by applying AI, machine learning and computer vision.
What is Alris?
AIris uses advanced image recognition technology to analyze the surrounding environment through the camera, offering intelligent suggestions for the most visually appealing compositions, optimal angles, and dynamic poses.
It provides real-time, actionable feedback, ensuring each shot is captured with the highest quality through continuous optimization.
Technology Utilize
ChatGPT
We use chatpgt to give suggestions on the modification of poses and other problems in pictures, and present them to users in the form of text and summary.
Google Teachable Machine
We trained a model using the google teachable machine to analyze the strength and weakness of light sources in a video or image in real time.
Ml5: PoseNet & NeuralNetwork Database
We used posenet's skeleton analysis function to track the user's shape and movement, and built a database to store different photo poses.
Working Protoype Video
Full Process
Video
Full Process
Video
Before Taking a Photo
Before taking photos, Alris will use the model summarized by machine learning to analyze factors such as the Angle of the light source. The upper right corner will directly display the situation of the light source, and the lower right corner will give corresponding suggestions.
Taking a Photo
After taking a photo, the AI will analyze whether the photo is suitable according to the trained model, and directly display the photo is good or bad in the upper left corner, and display detailed analysis of the reason below. If the photo is bad, a pose suggestion button appears at the bottom of the screen.
Pose Suggestion
In pose suggestion, the AI finds appropriate actions in the database and recommends them to the user based on the photo analysis results, and scans the user with computer vision. If the user makes this action correctly, it will automatically take a photo of the user and save it after a few seconds.