# 9/12/24 Meeting Notes #4

Classification:

Style Finder: Fine-Grained Clothing Style Detection and Retrieval: https://ieeexplore.ieee.org/document/6595844
- For specific clothing style, for the app to recommend outfits.
- Takes attributes from users as an input, AI will generate the outfit on the manniquinn.
-

# Two generative models recommended 推薦兩個生成模型

- Older model 舊型: GAN model (Generative Adversarial Networks ).
-- Can generate images faster. 可以更快地生成圖。

- Newer model 新型: Diffusion model.
-- Better for generative image as they can produce more detailed images. 對於生成圖像來說更好，因為它們可以產生更詳細的圖像。
-- Much easier to train compared to GAN models. 比起 GAN model 會跟容易訓練。

# Next week’s meeting outline下週會議：

- Survey papers on the two models and do a report on it, to figure out which model is more suitable to for our project. 看哪種模型更適合我們的專案。
- Draw out a flow chart and structure of what we do. (Two weeks deadline) 畫出流程圖/架構圖 （兩週截止日期）
- Figure out how the entire system works, how each part works, how it fixes the problems, receives and transfers data. Each model has different data, we need to figure out how to transfer the data to the next model. 弄清楚整個系統的運作方式、每個部分如何操作、如何解決問題、接收和傳輸數據。因為每個模型各有不同的數據，需要弄清楚如何把數據傳輸到下一個模型。
- Divide our work so that its more efficient.
- Introduce any papers that are related to our project and the models. We need to discuss the paper (such as the methods and models, not just the idea of the paper).

**Next meeting: September 19th 8pm in person.**

---
Previous: [8/29 Meeting Notes #3](https://hackmd.io/@emps-113up/meeting3)
Next: [9/19 Meeting Notes #5](https://hackmd.io/@emps-113up/meeting5)
Full Content List [here](https://hackmd.io/@emps-113up/full-list)