Datasets:

Modalities:
Image
Text
ArXiv:
DOI:
Libraries:
Datasets
License:
File size: 5,315 Bytes
d5c7af3
 
 
 
 
 
 
 
 
 
 
f405db9
d5c7af3
 
f405db9
 
 
 
 
 
 
 
 
 
d5c7af3
 
 
 
 
c438e68
 
4ab647e
839c582
c438e68
 
d5c7af3
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
083feca
 
 
45e702c
 
 
 
 
 
 
 
 
 
 
 
 
 
083feca
 
 
d5c7af3
 
 
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
---
license: cc-by-nc-4.0
task_categories:
- image-classification
tags:
- biology
- chemistry
size_categories:
- 1K<n<10K
---
# MobileMold: A Smartphone-Based Microscopy Dataset for Food Mold Detection

A smartphone-microsope-based dataset with 4941 annotated images for food mold detection

<div align="left">

  <a href="https://mobilemold.github.io/dataset/"><img src="https://img.shields.io/static/v1?label=Website&message=MobileMold&color=purple"></a> &ensp;
  <a href="https://doi.org/10.1145/3793853.3799806"><img src="https://img.shields.io/static/v1?label=Paper&message=ACM%20MMSys&color=black"></a> &ensp;
  <a href="https://arxiv.org/abs/2603.01944"><img src="https://img.shields.io/static/v1?label=Preprint&message=arXiv&color=red&logo=arxiv"></a> &ensp;
  <a href="https://huggingface.co/datasets/nphamdinh/mobilemold"><img src="https://img.shields.io/static/v1?label=Dataset&message=HuggingFace&color=yellow"></a> &ensp;
  <a href="https://www.kaggle.com/datasets/namphamdinh/mobilemold"><img src="https://img.shields.io/static/v1?label=Dataset&message=Kaggle&color=blue"></a> &ensp;
  <a href="https://doi.org/10.5281/zenodo.18782230"><img src="https://img.shields.io/static/v1?label=Dataset&message=Zenodo&color=green"></a> &ensp;
</div>

## 🌟 About MobileMold

**MobileMold** is a comprehensive dataset comprising **4,941 annotated images** for food mold detection, captured using smartphones with various clip-on microscope attachments.
The dataset addresses the growing need for accessible, low-cost food safety monitoring by leveraging smartphone-based microscopy. This enables research and development in computer vision applications for mold detection on various food surfaces.

[ACM MMSys] Paper: https://doi.org/10.1145/3793853.3799806 

Preprint:  https://arxiv.org/abs/2603.01944

Project Page: https://mobilemold.github.io/dataset/

---

### 📊 Dataset Overview
- **Total Images:** 4,941
- **Annotations:** Food Type and Mold Label
- **Food Types:** 11 categories (carrot, orange, creamcheese, tomato, toast, raspberry, mixed bread, blackberry, blueberry, cheese, onion)
- **Microscope Types:** 3 different clip-on smartphone microscopes (30x-100x magnification)
- **Smartphones:** Images captured with 3 different smartphone models
---

### 📁 Dataset Structure
  ```
MobileMold/
├── metadata.csv # Complete dataset metadata (4,941 entries)
├── train_metadata.csv # Training split metadata
├── val_metadata.csv # Validation split metadata
├── test_metadata.csv # Test split metadata
├── original/ # Original microscope images (as captured)
│ ├── L10 - 48.jpeg
│ ├── L10 - 25.jpeg
│ ├── L10 - 161.jpeg
│ └── ... (4,941 files total)
└── cropped_resized/ # Preprocessed images (same filenames)
├── L10 - 48.jpeg # Cropped to mold region & resized
├── L10 - 25.jpeg
├── L10 - 161.jpeg
└── ... (4,941 files, 1:1 mapping to original/)
  ```
---
### 📊 Dataset Composition

### Image Versions
1. **`original/`** - Raw images as captured by smartphone microscopes
   - Various resolutions (depending on smartphone and microscope)
   - Full field-of-view including background
   - Unprocessed image data

2. **`cropped_resized/`** - Processed images
   - Cropped to focus on mold regions
   - Resized to consistent dimensions
   - Same filenames as original folder

### Metadata Format
Each CSV file contains the following columns:

| Column | Description | Values/Examples |
|--------|-------------|-----------------|
| `filename` | Image filename (same in both folders) | `L10 - 48.jpeg` |
| `mold` | Binary indicator of mold presence | `True` / `False` |
| `food` | Type of food in image | `carrot`, `bread`, `cheese`, `tomato`, etc. |
| `phone` | Smartphone model used | `iPhone SE 2nd Generation`, etc. |
| `microscope` | Clip-on microscope model | `Apexel 100x`, etc. |

**Example metadata entry:**
```csv
filename,mold,food,phone,microscope
L10 - 48.jpeg,True,carrot,iPhone SE 2nd Generation,Apexel 100x
```

## FreshLens Mobile App

The [freshlens-app](https://github.com/MobileMold/freshlens-app) repository contains a Flutter-based mobile app designed for consumer-facing demonstrations and can be used in conjunction with a hosted model. Using a smartphone microscope attachment, users can capture or import images of food. The app then displays the probability that the food is moldy.

## Citation
If you use this useful for your research, please cite this as:
```
@inproceedings{Pham2026MobileMold,
  author    = {Pham, Dinh Nam and Prokisch, Leonard and Meyer, Bennet and Thumbs, Jonas},
  title     = {MobileMold: A Smartphone-Based Microscopy Dataset for Food Mold Detection},
  year      = {2026},
  isbn      = {9798400724817},
  publisher = {Association for Computing Machinery},
  address   = {New York, NY, USA},
  url       = {https://doi.org/10.1145/3793853.3799806},
  doi       = {10.1145/3793853.3799806},
  booktitle = {Proceedings of the ACM Multimedia Systems Conference 2026},
  pages     = {402--408},
  numpages  = {7},
  keywords  = {Dataset, Smartphone, Food, Mold, Microscope, Mobile, Fungal},
  series    = {MMSys '26}
}
```

## 📄 License

This dataset is available under the terms of the **[CC BY-NC 4.0](https://creativecommons.org/licenses/by-nc/4.0/)**