Dark Mode
Fruit Detection Image Set
Food & Beverage Consumption
Tags and Keywords
Trusted By
"No reviews yet"
Free
About
Fruit detection and classification tasks using machine learning models. It provides a rich collection of fruit images, pre-processed and augmented to enhance their utility for training and evaluating computer vision algorithms. The dataset supports multi-class classification, distinguishing between nine distinct fruit types. It is particularly well-suited for developing and refining object detection systems like YOLO.
Columns
The dataset, when presented in a tabular format such as a CSV, typically includes the following columns:
- filename: A unique identifier for each image file.
- Apple: A binary indicator (1 if an apple is present in the image, 0 otherwise).
- Banana: A binary indicator (1 if a banana is present in the image, 0 otherwise).
- Grapes: A binary indicator (1 if grapes are present in the image, 0 otherwise).
- Kiwi: A binary indicator (1 if a kiwi is present in the image, 0 otherwise).
- Mango: A binary indicator (1 if a mango is present in the image, 0 otherwise).
- Orange: A binary indicator (1 if an orange is present in the image, 0 otherwise).
- Pineapple: A binary indicator (1 if a pineapple is present in the image, 0 otherwise).
- Sugerapple: A binary indicator (1 if a sugerapple is present in the image, 0 otherwise).
- Watermelon: A binary indicator (1 if a watermelon is present in the image, 0 otherwise).
Distribution
The dataset comprises 2974 images in total. Each image has undergone auto-orientation of pixel data, including EXIF-orientation stripping, and has been resized to 640x640 pixels with stretching. To expand the dataset and improve model robustness, three augmented versions of each source image were created. Augmentation techniques applied include a 50% probability of horizontal flip, a 50% probability of vertical flip, random rotation between -15 and +15 degrees, random shear between -15° and +15° horizontally and vertically, and random Gaussian blur between 0 and 3.5 pixels. The dataset is structured for machine learning workflows, with defined paths for training, validation, and testing image sets.
Usage
This dataset is ideal for a variety of applications and use cases within computer vision, including:
- Developing and training object detection models for fruit recognition.
- Advancing research in image classification and machine learning.
- Creating automated systems for quality control in agriculture or retail.
- Educational purposes, demonstrating principles of data augmentation and model training.
- Benchmarking the performance of different YOLO versions or other convolutional neural networks on fruit detection tasks.
Coverage
The dataset focuses exclusively on images of nine common fruit types: Apple, Banana, Grapes, Kiwi, Mango, Orange, Pineapple, Sugerapple, and Watermelon. There is no specific geographic or time-range scope detailed for the image collection, suggesting a general applicability of the fruit images.
License
Attribution 4.0 International (CC BY 4.0)
Who Can Use It
This dataset is suitable for:
- Machine Learning Engineers: For training and evaluating object detection and classification models.
- Researchers: In the fields of computer vision, artificial intelligence, and agricultural technology.
- Students: Learning about image processing, deep learning, and dataset augmentation.
- Developers: Building applications that require automated fruit identification.
- Data Scientists: Exploring and experimenting with image-based datasets.
Dataset Name Suggestions
- Fruit Detection Image Set
- YOLO Fruit Image Classifier
- Augmented Fruit Vision Data
- Multi-Fruit Object Recognition
- Fresh Produce Image Collection
Attributes
Original Data Source: Fruit Detection Image Set
QUALITY