abhisheks008 · pacificrm · Oct 31, 2024 · Nov 6, 2024 · Nov 6, 2024 · Nov 6, 2024
diff --git a/Skin Disease Classification using Deep Learning Models/Dataset/datasetLink.txt b/Skin Disease Classification using Deep Learning Models/Dataset/datasetLink.txt
@@ -0,0 +1 @@
+Dataset : https://www.kaggle.com/datasets/pacificrm/skindiseasedataset/data
diff --git a/Skin Disease Classification using Deep Learning Models/Images/accuracy_plot.png b/Skin Disease Classification using Deep Learning Models/Images/accuracy_plot.png
diff --git a/...Disease Classification using Deep Learning Models/Images/class_distribution.png b/...Disease Classification using Deep Learning Models/Images/class_distribution.png
diff --git a/...Disease Classification using Deep Learning Models/Images/data_visualization.png b/...Disease Classification using Deep Learning Models/Images/data_visualization.png
diff --git a/Skin Disease Classification using Deep Learning Models/Images/loss_plot.png b/Skin Disease Classification using Deep Learning Models/Images/loss_plot.png
diff --git a/...Disease Classification using Deep Learning Models/Images/top2_accuracy_plot.png b/...Disease Classification using Deep Learning Models/Images/top2_accuracy_plot.png
diff --git a/Skin Disease Classification using Deep Learning Models/Model/Readme.md b/Skin Disease Classification using Deep Learning Models/Model/Readme.md
@@ -0,0 +1,98 @@
+# 🧴 Skin Disease Classification with Deep Learning 🧴
+
+> A comparative analysis of 10 different deep learning models for skin disease classification.
+
+## 🎯 Goal
+
+The aim of this project is to classify images of skin diseases into different categories by comparing various state-of-the-art deep learning architectures. We evaluated ten models, including transfer learning with pre-trained ImageNet weights, and identified the best-performing model for this specific task.
+
+## 📊 Dataset
+
+The dataset used for this project can be found on Kaggle: [Skin Disease Dataset](https://www.kaggle.com/datasets/pacificrm/skindiseasedataset/data). This dataset contains images of different types of skin diseases, providing a valuable resource for training and evaluating the model's performance.
+
+### 🌐 Comprehensive Collection
+
+This dataset comprises a diverse collection of images representing various skin diseases. The dataset images are categorized into 22 distinct classes, each corresponding to a specific skin condition.
+
+### 🌈 Diverse Skin Conditions
+
+The following are the classes included in the dataset:
+
+- **Acne**
+- **Actinic Keratosis**
+- **Benign Tumors**
+- **Bullous**
+- **Candidiasis**
+- **Drug Eruption**
+- **Eczema**
+- **Infestations/Bites**
+- **Lichen**
+- **Lupus**
+- **Moles**
+- **Psoriasis**
+- **Rosacea**
+- **Seborrheic Keratoses**
+- **Skin Cancer**
+- **Sun/Sunlight Damage**
+- **Tinea**
+- **Unknown/Normal**
+- **Vascular Tumors**
+- **Vasculitis**
+- **Vitiligo**
+- **Warts**
+
+### 📷 Data Visualization and Class Distribution
+
+We visualized the data to understand the sample characteristics for each class of skin disease. The following plots illustrate sample images and the class distribution:
+
+1. **Data Visualization**
+   ![Data Visualization](../Images/data_visualization.png)
+
+2. **Class Distribution**
+   ![Class Distribution](../Images/class_distribution.png)
+
+## 🧠 Models Used
+
+In this project, we evaluated the following ten deep learning architectures:
+
+1. **VGG16**
+2. **ResNet**
+3. **InceptionV3**
+4. **Xception**
+5. **InceptionResNetV2**
+6. **MobileNet**
+7. **ResNet50**
+8. **SEResNet**
+9. **EfficientNet**
+10. **Vision Transformer (ViT)**
+
+Each model (except SEResNet) utilized pre-trained weights from ImageNet, which provided a strong starting point for training. We observed that SEResNet, due to its lack of pre-training, performed significantly worse compared to other models. Among the models evaluated, **Vision Transformer (ViT)** yielded the best results in terms of both accuracy and generalization.
+
+## 📈 Results
+
+### 🔍 Performance Metrics
+
+We compared each model on the following metrics:
+
+- **Loss**: Tracking the loss during training and validation helped identify the stability of each model's learning process.
+- **Accuracy**: The primary metric used to gauge model performance.
+- **Top-2 Accuracy**: Provided insight into cases where the model's second choice was also relevant.
+
+The following plots provide a comparative visualization for each model's performance:
+
+1. **Loss**
+   ![Loss Plot](../Images/loss_plot.png)
+2. **Accuracy**
+   ![Accuracy Plot](../Images/accuracy_plot.png)
+3. **Top-2 Accuracy**
+   ![Top-2 Accuracy Plot](../Images/top2_accuracy_plot.png)
+
+## 🚀 How to Run
+
+To replicate this analysis, clone the repository and follow these steps:
+
+1. Install the required dependencies:
+   ```bash
+   pip install -r requirements.txt
+   ```
+2. Now you can run each cell of Notebook.
diff --git a/...Disease Classification using Deep Learning Models/Model/skin-disease-classification.ipynb b/...Disease Classification using Deep Learning Models/Model/skin-disease-classification.ipynb
diff --git a/Skin Disease Classification using Deep Learning Models/requirements.txt b/Skin Disease Classification using Deep Learning Models/requirements.txt
@@ -0,0 +1,9 @@
+tensorflow==2.12.0
+keras==2.12.0
+matplotlib==3.7.1
+numpy==1.23.5
+pandas==1.5.3
+scikit-learn==1.2.2
+seaborn==0.12.2
+opencv-python==4.7.0.72
+imghdr