CSCI 5117, Spring 2024, assignment description
- Team Name: 404notfound
- App Name: AI Art Museum
- App Link: https://project-1-404notfound.onrender.com/
- Jiyu Huang, [email protected]
- Xizhen Kang, [email protected]
- Zirui Chen, [email protected]
- Ruichen He, [email protected]
- Ziqi Zhou, [email protected]
Describe the most challenging features you implemented (one sentence per bullet, maximum 4 bullets):
- Personalized recommendation system
- image generated by openAI
Is there anything special we need to know in order to effectively test your app? (optional):
- recommendation by interaction and perference is updated 3 AM everyday. it could be manulally invoked by/api/recomtest for interaction recommendation and /api/get-recent-artworks/ for perference recommendation.
- for newly regidstered user, friends recommendation is not aviliable.
- when generating image throgh openAI api, the title must not include any special characters.
- this page provides user with 3 persionalized recommendation and 3 nonpersonalized recommendatiom. user can view, upload, search and go to profile form this page. _
- In this page, vistor can view the art work, and non-personalized recommendation will be provided.
- In this page, vistor can view the art work, including theautor, title, promopt and describtion about the art working, includind comments made by other user. visitor cannot make comments.
- this page show default login pafge provided by auth0
- this is to show detailed information about user.
- user can check out their fans and subscritons. they can click on it to see details.
- this page allow user to upload image they created or creat image based on their prompt on the fly.
- this page allow user to perfrom search. it will fuzzy search tiitle, prompt and details for a match
There are a few tools for mock-ups. Paper prototypes (low-tech, but effective and cheap), Digital picture edition software (gimp / photoshop / etc.), or dedicated tools like moqups.com (I'm calling out moqups here in particular since it seems to strike the best balance between "easy-to-use" and "wants your money" -- the free teir isn't perfect, but it should be sufficient for our needs with a little "creative layout" to get around the page-limit)
In this space please either provide images (around 4) showing your prototypes, OR, a link to an online hosted mock-up tool like moqups.com
Add images/photos that show your paper prototype (around 4) along with a very brief caption:
- In this page, user will choose either login to existing account or create a new account.Once users click login or register, they will be redirected to the Home Page.
- This is the Home Page of the website. On the top, it is the search bar that users can tap in key words that they want to search. On the right top corner, it is the button that user can use to navigate to User Profile Page. In the center of the page, it has a panel for Trending contents and a panel for recommendation contents (if user has login).In addition, for most of the pages, if the user has login, there will be a "+" button on the left bottom corner of the page, which users can use to navigate to Image Upload Page to upload their own art work. Clicking on any block (redpink/gray block, serve as place holder for art contents) will navigate to the Art Profile Page.
- This is the Image Upload page, where user uploads their images to the platform. User need to type in the title of the art, prompt that used to generate the art, and use the Upload button to select the art for upload.
- This the User Profile Page, which contains information about the user, which include Name, Description, Email, number of Subscription, Followers,and Like. It also has a list of art works. When user click any art work listed in the Profile page, they will be redirected to the Art Profile Page
- This is the key component of the platform, the Art Profile Page, which include title, description, prompt, author. It also provides the user option to follow the art work's author (if this page is accessed by a user that is not the owner of this art work). If the users have login, then they can also leave comments.
- These two images show the Search Bar, which exists in the majority of the pages of the platform. The search bar will include some reconmendation of the text fill (depends on if they previously visit the site, and if they login or not). By clicking "return", user will jump to a similar page with only the art contents that match the searching keyword.
Document integrations with 3rd Party code or services here. Please do not document required libraries. or libraries that are mentioned in the product requirements
- Library or service name: description of use
- AWS S3 for image storage
- tensorflow resNET and suprise for recommendation training
- openAI api for image generation
**If there's anything else you would like to disclose about how your project relied on external code, expertise, or anything else, please disclose that here: **
...