File size: 3,498 Bytes
954cb89 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 | ---
license: mit
configs:
- config_name: default
data_files:
- split: original
path: data/original-*
- split: augmented
path: data/augmented-*
dataset_info:
features:
- name: image
dtype: image
- name: label
dtype:
class_label:
names:
'0': food
'1': view
splits:
- name: original
num_bytes: 371487.0
num_examples: 32
- name: augmented
num_bytes: 4075054.0
num_examples: 352
download_size: 4411701
dataset_size: 4446541.0
---
# 📄 Model Card: Food vs View Image Dataset
## 1. Purpose
This dataset was created for educational purposes as part of Homework 1 (*Dealing with Data*).
It demonstrates how to build, preprocess, augment, and upload an image dataset to Hugging Face.
The dataset supports binary image classification tasks and shows how augmentation can expand small collections into larger, balanced sets.
---
## 2. Composition
- **Domain:** Everyday life photos.
- **Samples:** 32 unique, student-created images.
- **Classes (binary target):**
- `food` → meals, snacks, and beverages.
- `view` → landscapes, outdoor scenery.
- **Split Sizes:**
- Original: 32 images
- Augmented: 352 images
---
## 3. Collection Process
- **Source:** All original images were taken by the student using a phone camera.
- **Content:** Food photos are close-ups of meals, while view photos capture outdoor landscapes.
- **Format:** JPEG/PNG images.
- **Resolution:** Resized to 224×224 pixels.
- **Ethical Assurance:** No faces, people, or personal identifiers are included.
---
## 4. Preprocessing & Augmentation
- **Preprocessing:**
- Resized all images to 224×224.
- Converted to RGB format.
- **Augmentation Techniques (label-preserving):**
- Horizontal/vertical flips
- Random rotations (0°, 90°, 180°, 270°)
- Brightness/contrast adjustments
- Random cropping and scaling
- **Goal:** Expand dataset to ≥300 samples while preserving labels.
---
## 5. Labels
- **Target Variable:** `label`
- **Values:**
- `0`: food
- `1`: view
---
## 6. Splits
- **Original split:** 32 manually collected images (`original`).
- **Augmented split:** 352 synthetic variants via documented augmentation techniques (`augmented`).
---
## 7. Exploratory Data Analysis (EDA)
Below are random samples from both classes:
- **Class counts:** 15 food, 17 view
- **Observation:** Food images tend to be close-up shots under artificial lighting, while view images capture natural scenery with broader perspectives.
---
## 8. Intended Use / Limitations
- **Use:**
- Educational demo for binary image classification.
- Practice with preprocessing, augmentation, and Hugging Face uploads.
- **Limitations:**
- Small dataset; not suitable for real-world production.
- Augmentation does not fully capture real-world variability.
---
## 9. Ethical Considerations
- Images are student-owned and contain no personal or sensitive data.
- Dataset is safe for academic and research use.
---
## 10. License
Released under the **MIT License** for educational and research purposes.
---
## 11. AI Usage Disclosure
- **Original images (32):** Taken by the student.
- **Augmentation:** Generated via Python libraries (e.g., `torchvision.transforms`).
- **Documentation:** Drafted with the assistance of AI tools (ChatGPT) but reviewed and finalized manually.
|