NeuronZero/SkinCancerClassifier

SkinCancer-Classifier(small-sized model)

SkinCancer-Classifier is a fine-tuned version of swin-base. It was introduced in this paper by Liu et al. and first released in this repository. It was fine tuned on this dataset.

Model description

The Swin Transformer is a type of Vision Transformer. It builds hierarchical feature maps by merging image patches (shown in gray) in deeper layers and has linear computation complexity to input image size due to computation of self-attention only within each local window (shown in red). It can thus serve as a general-purpose backbone for both image classification and dense recognition tasks. In contrast, previous vision Transformers produce feature maps of a single low resolution and have quadratic computation complexity to input image size due to computation of self-attention globally.

Source

How to use

Here is how to use this model to identify melanoma from a picture of a the affected area of the skin:

from transformers import AutoImageProcessor, AutoModelForImageClassification
from PIL import Image
import requests

processor = AutoImageProcessor.from_pretrained("NeuronZero/SkinCancerClassifier")
model = AutoModelForImageClassification.from_pretrained("NeuronZero/SkinCancerClassifier")

# Dataset url: https://www.kaggle.com/datasets/nodoubttome/skin-cancer9-classesisic
 
image_url = "https://storage.googleapis.com/kagglesdsdata/datasets/319080/643971/Skin%20cancer%20ISIC%20The%20International%20Skin%20Imaging%20Collaboration/Test/melanoma/ISIC_0000049.jpg?X-Goog-Algorithm=GOOG4-RSA-SHA256&X-Goog-Credential=databundle-worker-v2%40kaggle-161607.iam.gserviceaccount.com%2F20240403%2Fauto%2Fstorage%2Fgoog4_request&X-Goog-Date=20240403T164047Z&X-Goog-Expires=345600&X-Goog-SignedHeaders=host&X-Goog-Signature=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"
image = Image.open(requests.get(image_url, stream=True).raw)

inputs = processor(images=image, return_tensors="pt")
outputs = model(**inputs)
logits = outputs.logits
predicted_class_idx = logits.argmax(-1).item()
print("Predicted class:", model.config.id2label[predicted_class_idx])