Merge pull request #870 from abckhush/main

Added Butterfly Image Classification
abhisheks008 · Jul 24, 2024 · 1fc8da3 · 1fc8da3
2 parents bca1cd7 + a16b07f
commit 1fc8da3
Show file tree

Hide file tree

Showing 10 changed files with 2,766 additions and 0 deletions.
diff --git a/Butterfly Image Classification/Dataset/README.md b/Butterfly Image Classification/Dataset/README.md
@@ -0,0 +1,5 @@
+# **Butterfly Image Classification**
+
+The dataset features 75 different classes of Butterflies. The dataset contains about 1000+ labelled images including the validation images. Each image belongs to only one butterfly category.
+
+### Dataset Link : https://www.kaggle.com/datasets/phucthaiv02/butterfly-image-classification/data
diff --git a/Butterfly Image Classification/Images/CNN Accuracy Plot.png b/Butterfly Image Classification/Images/CNN Accuracy Plot.png
diff --git a/Butterfly Image Classification/Images/Dataset.png b/Butterfly Image Classification/Images/Dataset.png
diff --git a/Butterfly Image Classification/Images/Distribution of Butterfly Classes.png b/Butterfly Image Classification/Images/Distribution of Butterfly Classes.png
diff --git a/Butterfly Image Classification/Images/EfficientNet Accuracy Plot.png b/Butterfly Image Classification/Images/EfficientNet Accuracy Plot.png
diff --git a/Butterfly Image Classification/Images/ResNet Accuracy Plot.png b/Butterfly Image Classification/Images/ResNet Accuracy Plot.png
diff --git a/Butterfly Image Classification/Models/butterfly-classification-cnn.ipynb b/Butterfly Image Classification/Models/butterfly-classification-cnn.ipynb
diff --git a/Butterfly Image Classification/Models/butterfly-classification-efficientnet.ipynb b/Butterfly Image Classification/Models/butterfly-classification-efficientnet.ipynb
diff --git a/Butterfly Image Classification/Models/butterfly-classification-resnet-50.ipynb b/Butterfly Image Classification/Models/butterfly-classification-resnet-50.ipynb
diff --git a/Butterfly Image Classification/README.md b/Butterfly Image Classification/README.md
@@ -0,0 +1,56 @@
+# **Butterfly Image Classification**
+
+## 🎯 Goal
+The primary goal of this project is to build and compare various deep learning models to accurately classify butterfly images into their respective species.
+
+## 🧵 Dataset : https://www.kaggle.com/datasets/phucthaiv02/butterfly-image-classification
+
+## 🧾 Description
+This dataset consists of over 1000 labeled images of butterflies, including validation images. Each image belongs to only one butterfly category. The challenge is to develop models that can accurately classify these images into the correct species.
+
+## 🚀 Models Implemented
+1. Convolutional Neural Network (CNN)
+2. EfficientNet
+3. ResNet50
+
+## 📚 Libraries Needed
+- TensorFlow: For building and training deep learning models.
+- Keras: For simplifying the creation and training of neural networks.
+- NumPy: For numerical computations and array operations.
+- Pandas: For data manipulation and analysis.
+- Matplotlib: For plotting and visualizing data.
+
+## 📊 Exploratory Data Analysis Results
+The dataset features 75 different classes of Butterflies. The dataset contains about 1000+ labelled images including the validation images. Each image belongs to only one butterfly category.
+![Distribution of Butterfly Classes](https://github.com/user-attachments/assets/b274368f-aa5e-4722-85c1-943339d36373)
+
+## 📈 Performance of the Models based on the Accuracy Scores
+
+**CNN Performance**
+
+![CNN Accuracy Plot](https://github.com/user-attachments/assets/e994354e-6693-4e38-87e6-948de0d1c524)
+
+**EfficientNet Performance**
+
+![EfficientNet Accuracy Plot](https://github.com/user-attachments/assets/0f0027ac-47d8-43f6-b4ba-6d9be0e8e876)
+
+**ResNet50 Performance**
+
+![ResNet Accuracy Plot](https://github.com/user-attachments/assets/60d54f34-bf63-4096-9bc4-01994927715e)
+
+## 📢 Conclusion
+The following models were implemented and evaluated based on their accuracy scores:
+
+### Accuracy Results
+| Model | Accuracy |
+|-------|----------|
+| CNN   |  87.36%  |
+| ResNet50 | 80.82% |
+| EfficientNet | 91.85%  |
+
+## Best Fitted Model
+EfficientNet achieved the highest accuracy of 91.85%, making it the best-performing model for this butterfly image classification task.
+
+## ✒️ Contributor
+- Name : Khushi Kalra
+- LinkedIn: https://www.linkedin.com/in/kalrakhushi/