Multi-Organism Bacterial Imaging Dataset
Description
The Bacterial Image Classification Dataset is a comprehensive collection of bacterial microscopy and culture images designed to support automated classification and CRISPR target identification research. The dataset encompasses 10 clinically relevant bacterial organisms, each represented through two distinct imaging modalities, resulting in 20 unique classification classes. Organisms The dataset includes the following 10 bacterial organisms: Aspergillus niger - A filamentous fungus commonly found in soil and food, used as a model organism for industrial applications Bacillus subtilis - A Gram-positive, rod-shaped bacterium, widely studied for its spore-forming capabilities and industrial applications Candida albicans - A yeast-like fungus and opportunistic pathogen, significant in clinical microbiology Clostridium sporogenes - A Gram-positive, spore-forming anaerobic bacterium, important in food safety and clinical diagnostics Enterococcus faecalis - A Gram-positive bacterium, clinically significant for its role in hospital-acquired infections and antibiotic resistance Escherichia coli - A Gram-negative rod-shaped bacterium, extensively studied model organism and indicator of fecal contamination Klebsiella pneumoniae - A Gram-negative bacterium, clinically important for its association with nosocomial infections and multidrug resistance Pseudomonas aeruginosa - A Gram-negative opportunistic pathogen, known for its resistance to antibiotics and biofilm formation Staphylococcus aureus - A Gram-positive coccus, clinically significant pathogen associated with various infections and antibiotic resistance Streptococcus pyogenes - A Gram-positive coccus, responsible for various human infections including strep throat and skin infections
Files
Steps to reproduce
Class Distribution The dataset maintains balanced representation across organisms, with slight variations in image counts per class: Aspergillus niger: 30 (gram stain) + 34 (media plate) = 64 images Bacillus subtilis: 28 (gram stain) + 29 (media plate) = 57 images Candida albicans: 30 (gram stain) + 29 (media plate) = 59 images Clostridium sporogenes: 29 (gram stain) + 26 (media plate) = 55 images Enterococcus faecalis: 27 (gram stain) + 26 (media plate) = 53 images Escherichia coli: 30 (gram stain) + 31 (media plate) = 61 images Klebsiella pneumoniae: 30 (gram stain) + 27 (media plate) = 57 images Pseudomonas aeruginosa: 27 (gram stain) + 33 (media plate) = 60 images Staphylococcus aureus: 53 (gram stain) + 38 (media plate) = 91 images Streptococcus pyogenes: 41 (gram stain) + 31 (media plate) = 72 images Dataset Characteristics Image Quality High-resolution digital captures ensuring clear morphological features Consistent imaging protocols across all samples Standardized preprocessing and normalization procedures Morphological Diversity Captures natural variations in bacterial morphology Includes variations in staining intensity and colony appearance Represents typical clinical and laboratory imaging conditions Clinical Relevance Organisms selected based on clinical significance Represents common pathogens encountered in clinical microbiology Supports research in antimicrobial resistance and CRISPR-based interventions Use Cases This dataset is designed to support: Automated Bacterial Classification Multi-class classification across 20 distinct classes Cross-modal learning (Gram stain ↔ Media plate) Morphological feature extraction and recognition CRISPR Target Identification Linking bacterial morphology to potential CRISPR targets Supporting multi-agent systems for gene target recommendation Enabling automated antimicrobial intervention design Antimicrobial Resistance (AMR) Profiling Associating morphological features with resistance patterns Supporting AMR prediction from image analysis Facilitating rapid resistance detection Biofilm Detection Identifying biofilm-associated morphological characteristics Supporting biofilm gene detection from images Enabling biofilm formation prediction Clinical Decision Support Rapid bacterial identification from microscopy images Supporting diagnostic workflows in clinical laboratories Enabling automated preliminary classification