Add README.md

Co-authored-by: SimoCimmi <simonecimmino2004@gmail.com>
Co-authored-by: Morgan Vitiello <morgan.vitiello06@gmail.com>

Author: 3 people

.github/images/baseline.png

@@ -0,0 +1,62 @@
+# SkinDetector
+SkinDetector is a deep-learning AI module capable of diagnosing canine dermatological diseases by observing an image of your dog's skin.
+
+SkinDetector is part of the [SafePet project](https://github.com/Progetto-SafePet).
+
+![SafePet Logo](./.github/images/safepet.png)
+
+
+## Authors
+Simone Cimmino [@SimoCimmi](https://github.com/SimoCimmi) <br>
+Luca Salvatore [@lucasalvaa](https://github.com/lucasalvaa) <br>
+Morgan Vitiello [@MorganVitiello](https://github.com/MorganVitiello)
+
+## Credits
+The project was developed at the University of Salerno, Department of Computer Science, in the academic year 2025-26 for
+the exam of Fundamentals of Artificial Intelligence helb by Professor Fabio Palomba [@fpalomba](https://github.com/fpalomba), whom we thank for his support.
+
+## Dataset
+The dataset on which the models were trained is [Dog's skin diseases (Image Dataset)](https://www.kaggle.com/datasets/youssefmohmmed/dogs-skin-diseases-image-dataset).
+
+![class distribution](./.github/images/classes.png)
+
+## Pipelines
+To select the final model, four slightly different training pipelines were undertaken. Each pipeline is defined in the experiments/[pipeline] directory in the relative dvc.yaml file.
+
+### Baseline
+The training will be performed on the starting dataset without duplicates.
+
+![baseline](./.github/images/baseline.png)
+
+### Data Augmentation (Pipeline 1)
+Model training will be performed after creating synthetic samples to increase 
+the size of the dataset. We selected four techniques to increase the image volume of the training set:
+horizontal flipping, vertical flipping, noise injection, and saturation enhancement. Of these four,
+only two techniques will be applied to each image in the initial training set, according to a
+randomized criterion. After the data augmentation phase, the size of the training set 
+will be tripled.
+
+![pipeline1](./.github/images/pipeline1.png)
+
+### Two-Stage Fine-Tuning (Pipeline 2)
+This approach, formalized by [ValizadehAslani et al.](https://arxiv.org/abs/2207.10858), is designed to address the
+class-imbalance problem of the training set. Unlike standard fine-tuning,
+which often causes the model to ignore minority classes in favor of majority classes,
+this technique divides the learning process into two distinct phases.
+In the first phase, only the last layer of the model is trained, freezing all other backbone layers. 
+The weighted loss function [LDAM Loss](https://arxiv.org/abs/1906.07413) is used, which assigns larger margins to minority classes, 
+forcing the classifier to minimize the error on them despite the sparseness of samples.
+In the second phase, fine-tuning is performed by unfreezing all model layers and using the traditional Cross-Entropy loss function.
+
+![pipeline2](./.github/images/pipeline2.png)
+
+
+### Data Augmentation + Two-Stage Fine-Tuning (Pipeline 3)
+We combine the data augmentation strategy of Pipeline 1 with the two-stage fine-tuning strategy of Pipeline 2.
+
+![pipeline3](./.github/images/pipeline3.png)
+
+## Results
+
+The table shows the results of the experiments. The final model chosen is EfficientNetV2_S trained in pipeline 3.
+![results](.github/images/results.png)