Labeled Mushroom images suitable for training and evaluating computer vision and deep learning models.

This dataset contains 2,820 high-quality images of various mushroom species, meticulously categorized into edible and poisonous classes. The dataset provides a diverse collection of 25 edible mushroom species and 22 poisonous mushroom species, capturing a wide range of visual features such as color, shape, texture, and size variations.

To ensure an optimal data distribution for machine learning and deep learning applications, the dataset has been strategically split into three subsets:

Training Set (80%) – Comprising 2,256 images, this subset serves as the primary source for model learning, offering a comprehensive range of mushroom species. Validation Set (10%) – Consisting of 282 images, this subset is used for hyperparameter tuning and performance evaluation to mitigate overfitting. Test Set (10%) – Including 282 images, this subset is reserved for final model evaluation to assess its generalization capability on unseen data. The dataset is structured into two main categories:

Edible Mushrooms (25 species) Poisonous Mushrooms (22 species) Unlike raw datasets that maintain species-specific subdirectories, this dataset organizes images directly under their respective labels (edible/ and poisonous/). This format simplifies its integration with deep learning frameworks such as TensorFlow, PyTorch, and Keras, making it well-suited for image classification tasks, mushroom species recognition, and toxicity prediction.

This dataset can be applied to various domains, including computer vision research, mycology studies, AI-driven foraging assistance, and food safety analysis.

Collection Credit: https://www.kaggle.com/datasets/yoonjunggyu/25-edible-mushroom-and-25-poisonous-mushroom

This dataset is used to classify 9 kinds of mushroom.
If you like this dataset and find it useful, please give a thumbs up 👍👍👍👍
💥💥💥💥💥
Pay attention that there are some truncated images in this dataset(maybe only one, I'm not sure), so you are not recommended to use function like ImageFolder to load the whole dataset.
Otherwise, you will get OS Error, which I have shown you in my notebook. It's better to rewrite Dataset class.

1) Data Introduction • The DeFungi: Microscopic Fungi Image Classification Dataset is a microscopic image dataset designed to classify various types of superficial fungal infections, including yeasts, moulds, and dermatophytes.

2) Data Utilization (1) Characteristics of the DeFungi: Microscopic Fungi Image Classification Dataset: • The dataset contains microscopic visual information of fungal infections, allowing models to learn the cellular structure and morphological differences of pathogens. • It is specialized in superficial fungal infections and provides a data environment that closely reflects real-world diagnostic scenarios.

(2) Applications of the DeFungi: Microscopic Fungi Image Classification Dataset: • Development of fungal infection classification models: Can be used to train deep learning models that automatically classify fungal types such as yeasts, moulds, and dermatophytes. • Research on AI-assisted diagnostic tools: Suitable for developing AI-based systems that quickly detect and categorize infection types using medical imaging.

Context

We created this dataset as part of our school's research project. As we didn't find something similar when we started, we decided to publish it here so that future research with mushrooms and AI can benefit from it.

Content

The dataset is separated in 2 classes , mushroom sporocarps and not mushroom sporocarps. Each class has 2 subclasses, edible and non-edible(includes medicinal, unpalatable and hallucinogenic fungi).

Acknowledgements

A HUGE thanks to my colleagues: Gabriel and Bianca. Without you guys this dataset wouldn't be possible

Inspiration

(◕ᴗ◕✿) Be cool!

1) Data Introduction • Edible & Poisonous Fungi – Fungus vs. Mushroom Dataset is a binary classification image dataset created to distinguish whether an image depicts a non-mushroom sporocarp (fungus) or a typical mushroom.

2) Data Utilization (1) Characteristics of the Edible & Poisonous Fungi – Fungus vs. Mushroom Dataset: • This dataset contains both non-mushroom sporocarp (fungus) images and mushroom images, making it suitable for binary classification tasks that clearly distinguish between the two content groups.

(2) Applications of the Edible & Poisonous Fungi – Fungus vs. Mushroom Dataset: • Content Classification Model Development: This dataset can be used to train AI-based binary classifiers that automatically distinguish between non-mushroom sporocarps (fungi) and mushrooms. The trained model can be deployed on embedded field survey devices or mobile apps for real-time image filtering and related applications.

1) Data Introduction • The Edible & Poisonous Fungi – Mushroom Sporocarp Edibility Dataset is a binary classification image dataset composed exclusively of mushroom sporocarp images, created to distinguish whether each mushroom sporocarp is edible or poisonous.

2) Data Utilization (1) Characteristics of the Edible & Poisonous Fungi – Mushroom Sporocarp Edibility Dataset: • Since it contains only mushroom sporocarp images, this dataset is optimized for binary classification tasks focused on mushroom edibility classification.

(2) Applications of the Edible & Poisonous Fungi – Mushroom Sporocarp Edibility Dataset: • Edibility Classification Model Development: This dataset can be used to train AI-based binary classifiers that take mushroom sporocarp images as input and accurately determine edibility. • Safety & Educational Applications: The dataset can be applied to implement real-time edibility warning features in field survey tools or mushroom foraging guide apps.

Context

This dataset contains 3122 images (512x512) of 215 different mushrooms (at least 4 images per class). mushrooms.txt contains names of all the mushrooms in the dataset.

Legal notice

The dataset was created with the images, that were obtained from WildFoodUK. I'm not responsible for the use of the dataset

Inspiration

The original idea was to create an android application with mushrooms detection capabilities, but all that was done is this dataset.

Mushroom Classification

## Overview

Mushroom Classification is a dataset for classification tasks - it contains Mushroom annotations for 258 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

  ## License

  This dataset is available under the [CC BY 4.0 license](https://creativecommons.org/licenses/CC BY 4.0).

Description:

This dataset comprises images of 12 distinct species of mushrooms, collected from various sources, primarily sourced from Mushroom World. The species included are Agaricus, Amanita, Boletus, Cortinarius, Entoloma, Exidia, Hygrocybe, Inocybe, Lactarius, Pluteus, Russula, and Suillus, representing a diverse array of fungal taxa.

Dataset Contents: - Images of mushrooms from each species, captured under varying conditions. - Ground truth labels are provided for each image, indicating the species of mushroom depicted. - Images have been processed using GrabCut segmentation to isolate the mushrooms from their backgrounds, facilitating easier analysis and classification.

Applications:

• Classification and identification of mushroom species using machine learning and computer vision techniques.
• Research into fungal taxonomy and biodiversity.
• Educational purposes, including mushroom identification guides and apps.

Usage Notes:

• This dataset is suitable for training and testing machine learning models for mushroom species classification.
• Researchers and enthusiasts interested in mycology can utilize this dataset for various analyses and studies.
• Proper citation of the original data sources, including Mushroom World, is encouraged when using this dataset in publications or projects.

Acknowledgments:

• The source of the images is primarily from Mushroom World, and credit is given to them for providing the foundational data for this dataset.
• The GrabCut segmentation technique used to process the images was applied to isolate the mushrooms from their backgrounds, enhancing the usability of the dataset for classification tasks.

Note to Users: - Please ensure compliance with any licensing or usage restrictions associated with the original images obtained from Mushroom World. - Feedback and contributions to enhance the dataset are welcome and appreciated.

Published Paper similar to this dataset: https://www.hindawi.com/journals/jfq/2022/1173102/

Mushroom Disease Detection

## Overview

Mushroom Disease Detection is a dataset for object detection tasks - it contains Healthy Affected annotations for 56 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

  ## License

  This dataset is available under the [CC BY 4.0 license](https://creativecommons.org/licenses/CC BY 4.0).

The dataset includes 8857 images. Mushroom are annotated in COCO format.

The following pre-processing was applied to each image: * Auto-orientation of pixel data (with EXIF-orientation stripping) * Resize to 640x640 (Stretch)

The following augmentation was applied to create 3 versions of each source image: * 50% probability of horizontal flip * 50% probability of vertical flip

The structure:

dataset-directory/
├─ README.dataset.txt
├─ README.roboflow.txt
├─ train
│ ├─ train-image-1.jpg
│ ├─ train-image-1.jpg
│ ├─ ...
│ └─ _annotations.coco.json
├─ test
│ ├─ test-image-1.jpg
│ ├─ test-image-1.jpg
│ ├─ ...
│ └─ _annotations.coco.json
└─ valid
  ├─ valid-image-1.jpg
  ├─ valid-image-1.jpg
  ├─ ...
  └─ _annotations.coco.json

To convert the format to YOLO annotations, go to roboflow.

Mushroom cultivation bags, categorized into healthy, single-infected, and mixed-infected classes, were photographed at the Mushroom Development Institute in Savar, Dhaka, Bangladesh. The dataset comprises a total of 680 high-resolution images, including 299 healthy samples, 72 single-infected ones, and 309 mixed-infected images. The infected samples are contaminated with green mold, black mold, or both. The mixed-infected category represents bags affected by multiple pathogens or overlapping infection patterns, typically involving a combination of these two types of mold.

Grey Oyster Mushroom Grading

## Overview

Grey Oyster Mushroom Grading is a dataset for classification tasks - it contains Mushroom annotations for 2,600 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

  ## License

  This dataset is available under the [CC BY 4.0 license](https://creativecommons.org/licenses/CC BY 4.0).

Mushroom image dataset for classification, based on mushrooms in Bosnia. Sources: mushroom species pulled from "Gljive: šumsko bogatstvo Bosne i Hercegovine" book by Hajrudin Rudi Hasanbegović. Mushroom images for dataset based on https://www.kaggle.com/datasets/derekkunowilliams/mushrooms and https://github.com/bechtle/mushroomobser-dataset

The file one_million_mushrooms.csv consists of just over 1 million rows of synthetic data generated form the Secondary data GitHub repo.

There is a second file, test.csv which contains another 0.5 million rows (note: also contains the target column class) which can be used as a hold-out validation set.

Related datasets on Kaggle: * UCI Mushroom Classification - the original dataset of 8124 real mushrooms * Secondary Mushroom - 61,069 synthetic mushrooms

Credit: dataset image made using Flux.1 AI.

The automatic identification of edible fungus based on deep learning can provide technical support for automatic marking and pricing for unmanned supermarkets and vegetable markets in the construction of smart city. Through the identification of intelligent traceability scale, the type and price of edible fungus are automatically displayed, which can reduce the consumption of human resources and save time and cost. At present, the machine recognition of edible fungus mainly depends on a small number of edible fungi pictures collected by some researchers independently in the experimental environment, and there is a lack of edible fungus picture samples obtained in the complex natural environment. This YMushroom dataset provides high-definition edible fungus images that can be used for deep learning image classification model training, including dry and fresh edible fungus in different seasons, different acquisition backgrounds and different acquisition equipment. The dataset is divided into 28 categories, with a total of 49958 pictures. Among them, the sample size of Shaggy Cap pictures is the least, and Oyster Mushroom pictures is the most, 969 and 2578 respectively. The median sample size of a single edible fungus type is 1764, which can meet the training needs of mainstream deep learning models. This dataset can provide basic data for edible fungus image classification, object detection, semantic segmentation, panoptic segmentation and other research.

MushR General Summary: MushR is a modular and scalable gourmet mushroom growing and harvesting system that goes beyond the state of the art, which merely monitors and controls the growing environment, by introducing an image recognition system that determines when and which mushrooms are ready to be harvested in conjunction with a proof-of-concept of an automated mushroom harvesting mechanism for harvesting the mushrooms without human interaction. The image recognition setup monitors the growing status of the mushrooms and guides the harvesting process. We present a Mask R-CNN model for the detection of oyster mushroom maturity as well as a semi-automated harvesting system, integrating a Raspberry Pi for control, an electrical switch, an air compressor, and a pneumatic cylinder with a cutting knife to facilitate timely mushroom harvesting. The modularity and scalability of the system allow for industry-level usage and can be scaled according to the required mushroom-growing systems within the facility. MushR Dataset: The dataset created for this project focuses on capturing images of the mushroom-growing environment from three different perspectives within each of our two growth tents for mushroom production. Instead of providing images of every individual bucket and mushroom, we capture the overall scene and its variations. The images from each perspective are captured simultaneously and automatically hourly. This approach allows for monitoring the development and maturity of the oyster mushrooms over time. We captured and accumulated 34,400 images over ten months to ensure a comprehensive dataset.

Mushroom_classification_v2

## Overview

Mushroom_classification_v2 is a dataset for classification tasks - it contains Mushrooms ZO51 annotations for 901 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

  ## License

  This dataset is available under the [CC BY 4.0 license](https://creativecommons.org/licenses/CC BY 4.0).

Mushroom_classification_v3

## Overview

Mushroom_classification_v3 is a dataset for classification tasks - it contains Mushrooms D3mW annotations for 1,248 images.

## Getting Started

You can download this dataset for use within your own projects, or fork it into a workspace on Roboflow to create your own model.

  ## License

  This dataset is available under the [CC BY 4.0 license](https://creativecommons.org/licenses/CC BY 4.0).