Thomasbehan / LesNet

LesNet (Lesion Net) is an open-source project for AI-based skin lesion detection. It aims to create a reliable tool and foster community involvement in critical AI problems. Contributions are welcome!
Mozilla Public License 2.0
12 stars 2 forks source link

Training the model with custom data #27

Closed upstream-team closed 11 months ago

upstream-team commented 1 year ago

I have a question regarding machine learning. I am wondering if it is possible to train a system using custom data to accurately detect and classify skin problems such as rashes and acne. I am specifically interested in identifying the type of rashes and acne. I would appreciate any insights or guidance on how to approach this. Thank you!

Thomasbehan commented 11 months ago

Absolutely, you can extend the capability of SkinvestigatorAI to detect and classify skin problems like rashes and acne. Just add your images to the /data directory (if there isn't one you can create one). Organize them in test, train, and validation subfolders. Within each of these subfolders, create a separate folder for each diagnosis you want the AI to detect. For instance: /data/validation/melanoma/image.png. or /data/train/eczema/image.jpg You can store as many images in each diagnoses folder as you like.

Once your data is organized, you can train the AI by running the training script: /skinvestigatorai/core/ai/train.py.

Hope this helps.