House Plant Species Dataset

Description:

This comprehensive dataset contains a total of 14,790 images, each carefully categorized into one of 47 plant species. Originally curated for a project focused on determining whether various plant species are toxic or safe for cats, the dataset serves as a valuable resource for plant classification models. While the images were sourced from Bing, the data has been manually curated by non-experts, ensuring a balanced and diverse range of species, although the dataset’s accuracy may benefit from further refinement by botanical experts.

Download Dataset

Key Features:

Total Images: 14,790

Species Categories: 47 distinct plant types

Image Source: Bing Images

Curation: Non-expert manual cleaning

Dataset Breakdown:

Class sizes vary; popular species include Monstera Deliciosa (547 images), Dumb Cane (541 images), and Chinese Evergreen (514 images).

Species with fewer representations include Yucca (66 images), Kalanchoe (130 images), and Asparagus Fern (169 images).

Image Characteristics:

Quality & Resolution: Images in the dataset exhibit a range of resolutions, with some variance in quality due to the nature of the collection process. Despite these inconsistencies, the variety offers opportunities for models to learn from different image qualities.

Types of Images: The dataset includes both whole-plant images and detailed close-ups of plant parts, providing multi-dimensional perspectives of each species.

Environmental Diversity: Plants are photographed in various settings, both indoors and outdoors, to account for different visual contexts.

Organized Structure: Each species is neatly organized into its respective folder, allowing for easy access and classification.

This dataset is sourced from Kaggle.