Update README.md

README.md

@@ -12,63 +12,23 @@ tags:
 - fiftyone
 - image
 - image-classification
+- transfer-learning
+- vgg16
+- binary-classification
+- computer-vision
+- pets
+- dogs
 overwrite: true
-dataset_summary: '
-
-
-
-
-  This is a [FiftyOne](https://github.com/voxel51/fiftyone) dataset with 169 samples.
-
-
-  ## Installation
-
-
-  If you haven''t already, install FiftyOne:
-
-
-  ```bash
-
-  pip install -U fiftyone
-
-  ```
-
-
-  ## Usage
-
-
-  ```python
-
-  import fiftyone as fo
-
-  from fiftyone.utils.huggingface import load_from_hub
-
-
-  # Load the dataset
-
-  # Note: other available arguments include ''max_samples'', etc
-
-  dataset = load_from_hub("andandandand/bo_or_not")
-
-
-  # Launch the App
-
-  session = fo.launch_app(dataset)
-
-  ```
-
-  '
+dataset_summary: 'Binary classification dataset for identifying Bo (Barack Obama''s Portuguese Water Dog) versus other pets. This FiftyOne dataset contains 169 samples and was created for demonstrating transfer learning with VGG16.'
 ---
 
-# Dataset Card for bo-dataset
-
-<!-- Provide a quick summary of the dataset. -->
-
-
+![Bo Dataset](bo_dataset.png)
 
+# Dataset Card for bo-dataset
 
+[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/drive/1XFoKgM_WQ9l2WgK6aS5GLFGc8uv0cdGB)
 
-This is a [FiftyOne](https://github.com/voxel51/fiftyone) dataset with 169 samples.
+This is a [FiftyOne](https://github.com/voxel51/fiftyone) dataset with 169 samples designed for binary classification of Bo (Barack Obama's Portuguese Water Dog) versus other pets.
 
 ## Installation
 
@@ -92,135 +52,161 @@ dataset = load_from_hub("andandandand/bo_or_not")
 session = fo.launch_app(dataset)
 ```
 
-
 ## Dataset Details
 
 ### Dataset Description
 
-<!-- Provide a longer summary of what this dataset is. -->
-
+This dataset contains images for binary classification between Bo (Barack Obama's Portuguese Water Dog) and other pets (cats and dogs). The dataset was created to demonstrate transfer learning techniques using a pre-trained VGG16 model. Bo was a Portuguese Water Dog who lived in the White House during Barack Obama's presidency.
 
+- **Curated by:** Antonio Rueda-Toicen ([email protected])
+- **Language(s):** en
+- **License:** Creative Commons Attribution 4.0 International License (CC BY 4.0)
+- **Repository:** [Colab Notebook](https://colab.research.google.com/drive/1XFoKgM_WQ9l2WgK6aS5GLFGc8uv0cdGB)
 
-- **Curated by:** [More Information Needed]
-- **Funded by [optional]:** [More Information Needed]
-- **Shared by [optional]:** [More Information Needed]
-- **Language(s) (NLP):** en
-- **License:** mit
+### Dataset Sources
 
-### Dataset Sources [optional]
-
-<!-- Provide the basic links for the dataset. -->
-
-- **Repository:** [More Information Needed]
-- **Paper [optional]:** [More Information Needed]
-- **Demo [optional]:** [More Information Needed]
+- **Original Data Source:** [Presidential Doggy Door - Kaggle](https://www.kaggle.com/datasets/drvnmanju/presidential-doggy-door)
+- **Tutorial Implementation:** [Google Colab Notebook](https://colab.research.google.com/drive/1XFoKgM_WQ9l2WgK6aS5GLFGc8uv0cdGB)
+- **Alternative Access:** [Google Drive](https://drive.google.com/drive/folders/1EU1ujeNtvsaKAIeyS7J8BDFC60iHvwi1?usp=drive_link)
 
 ## Uses
 
-<!-- Address questions around how the dataset is intended to be used. -->
-
 ### Direct Use
 
-<!-- This section describes suitable use cases for the dataset. -->
-
-[More Information Needed]
+This dataset is intended for:
+- Binary image classification tasks
+- Transfer learning demonstrations
+- Computer vision education and tutorials
+- Experimenting with pre-trained models like VGG16
+- Learning FiftyOne dataset management and visualization
 
 ### Out-of-Scope Use
 
-<!-- This section addresses misuse, malicious use, and uses that the dataset will not work well for. -->
-
-[More Information Needed]
+This dataset should not be used for:
+- Production security systems
+- Real-world pet identification systems
+- Commercial applications without proper validation
+- Any application requiring high accuracy pet identification
 
 ## Dataset Structure
 
-<!-- This section provides a description of the dataset fields, and additional information about the dataset structure such as criteria used to create the splits, relationships between data points, etc. -->
+The dataset contains 169 images split across three sets:
+- **Training set:** 50% of original training data
+- **Validation set:** 50% of original training data  
+- **Test set:** Independent test images
+
+### Data Fields
+
+- `filepath`: Path to the image file
+- `ground_truth`: Classification label ("bo" or "not_bo")
+- `tags`: Dataset split indicators ("train", "validation", "test")
+- `vgg16-imagenet-predictions`: Original VGG16 ImageNet predictions
+- `vgg16-imagenet-embeddings`: Feature embeddings from VGG16
+- `fine_tuned_vgg16_prediction`: Fine-tuned model predictions (test set only)
+
+### Label Distribution
 
-[More Information Needed]
+- **bo**: Images of Bo (Barack Obama's Portuguese Water Dog)
+- **not_bo**: Images of other pets (cats and dogs)
 
 ## Dataset Creation
 
 ### Curation Rationale
 
-<!-- Motivation for the creation of this dataset. -->
-
-[More Information Needed]
+This dataset was created to demonstrate transfer learning concepts using a real-world scenario where a computer vision system needs to identify a specific individual (Bo) among other similar animals. The task simulates a security application while providing an engaging educational example.
 
 ### Source Data
 
-<!-- This section describes the source data (e.g. news text and headlines, social media posts, translated sentences, ...). -->
-
 #### Data Collection and Processing
 
-<!-- This section describes the data collection and processing process such as data selection criteria, filtering and normalization methods, tools and libraries used, etc. -->
-
-[More Information Needed]
+The images were collected and organized into a binary classification structure:
+- Images of Bo were labeled as "bo"
+- Images of other pets (cats and dogs) were labeled as "not_bo"
+- Data was split into training/validation and test sets
+- Images were processed using standard computer vision preprocessing techniques
 
-#### Who are the source data producers?
-
-<!-- This section describes the people or systems who originally created the data. It should also include self-reported demographic or identity information for the source data creators if this information is available. -->
-
-[More Information Needed]
-
-### Annotations [optional]
+#### Data Augmentation
 
-<!-- If the dataset contains annotations which are not part of the initial data collection, use this section to describe them. -->
+The training process includes augmentation techniques:
+- Affine transformations (translation, scaling, rotation)
+- Elastic deformations
+- Perspective transformations
+- Horizontal flipping
+- Brightness and contrast adjustments
+- Random grayscale conversion
 
-#### Annotation process
+#### Who are the source data producers?
 
-<!-- This section describes the annotation process such as annotation tools used in the process, the amount of data annotated, annotation guidelines provided to the annotators, interannotator statistics, annotation validation, etc. -->
+The dataset was curated by Antonio Rueda-Toicen for educational purposes as part of FiftyOne documentation and tutorials.
 
-[More Information Needed]
+### Model Training Details
 
-#### Who are the annotators?
+The dataset includes results from transfer learning using:
+- **Base Model:** VGG16 pre-trained on ImageNet
+- **Architecture Modification:** Replaced final classifier for binary classification
+- **Training Strategy:** Froze base VGG16 layers, trained only new classifier layers
+- **Loss Function:** Binary Cross Entropy with Logits
+- **Optimizer:** Adam (lr=0.003)
+- **Training Epochs:** 10 epochs with early stopping
 
-<!-- This section describes the people or systems who created the annotations. -->
+## Technical Implementation
 
-[More Information Needed]
+### Preprocessing Pipeline
 
-#### Personal and Sensitive Information
+Images are processed using:
+- Conversion to RGB format
+- Resizing to 256x256 pixels
+- Center cropping to 224x224 pixels
+- Normalization with ImageNet statistics (mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
 
-<!-- State whether the dataset contains data that might be considered personal, sensitive, or private (e.g., data that reveals addresses, uniquely identifiable names or aliases, racial or ethnic origins, sexual orientations, religious beliefs, political opinions, financial or health data, etc.). If efforts were made to anonymize the data, describe the anonymization process. -->
+### Evaluation Metrics
 
-[More Information Needed]
+The model performance is evaluated using:
+- Binary classification accuracy
+- Confusion matrix
+- Per-class precision and recall
+- FiftyOne's built-in evaluation tools
 
 ## Bias, Risks, and Limitations
 
-<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+### Limitations
 
-[More Information Needed]
+- Small dataset size (169 samples) limits generalization
+- Limited to specific breeds and individuals
+- May not generalize to other Portuguese Water Dogs
+- Training data may not represent full diversity of pet appearances
+- Designed for educational purposes, not production use
 
-### Recommendations
+### Potential Biases
 
-<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+- Dataset may be biased toward specific lighting conditions, angles, or image quality
+- Limited representation of pet diversity
+- Potential overfitting due to small dataset size
 
-Users should be made aware of the risks, biases and limitations of the dataset. More information needed for further recommendations.
+### Recommendations
 
-## Citation [optional]
+Users should be aware that this dataset is primarily educational and should not be used for production applications without significant additional validation and testing. The small size makes it unsuitable for robust real-world applications.
 
-<!-- If there is a paper or blog post introducing the dataset, the APA and Bibtex information for that should go in this section. -->
+## Citation
 
 **BibTeX:**
-
-[More Information Needed]
+```bibtex
+@misc{rueda_toicen_2024_bo_dataset,
+  author = {Rueda-Toicen, Antonio},
+  title = {Bo or Not Bo: Binary Classification with Transfer Learning},
+  year = {2024},
+  url = {https://colab.research.google.com/drive/1XFoKgM_WQ9l2WgK6aS5GLFGc8uv0cdGB},
+  note = {FiftyOne educational dataset}
+}
+```
 
 **APA:**
+Rueda-Toicen, A. (2024). Bo or Not Bo: Binary Classification with Transfer Learning. FiftyOne Educational Dataset. https://colab.research.google.com/drive/1XFoKgM_WQ9l2WgK6aS5GLFGc8uv0cdGB
 
-[More Information Needed]
-
-## Glossary [optional]
-
-<!-- If relevant, include terms and calculations in this section that can help readers understand the dataset or dataset card. -->
-
-[More Information Needed]
-
-## More Information [optional]
-
-[More Information Needed]
-
-## Dataset Card Authors [optional]
+## Dataset Card Authors
 
-[More Information Needed]
+Antonio Rueda-Toicen ([email protected])
 
 ## Dataset Card Contact
 
-[More Information Needed]
+For questions about this dataset, please contact Antonio Rueda-Toicen at [email protected] or visit the [FiftyOne documentation](https://docs.voxel51.com/).