Decision-support system for live detection of Leishmania parasites from microscopic images with Deep Learning

1. Overview  

This dataset consists of microscopic images of Giemsa-stained skin smears obtained from patients diagnosed with cutaneous leishmaniasis (CL).  
It is organized into two main parts:  

• Dataset 1  → Collected with a Keyence BZ9000E digital microscope (lab-based).  
• Dataset 2  → Extended dataset including all images from Dataset 1, plus an additional set collected with a Bresser Erudit DLX microscope (portable, low-cost).  

Both datasets contain paired Images (.png) and Labels (.txt), split into train, val, and test subsets.

2. Data Acquisition

Dataset 1  

• Patients: 244 Libyan CL patients (confirmed by PCR at Tripoli University Hospital)
• Samples: Skin lesion smears (slit-skin or touch smears)
• Preparation: Air-dried, methanol-fixed, Giemsa-stained slides
• Imaging Setup:  
  • Microscope: Keyence BZ9000E (lab-grade)  
  • Magnification: 100× oil immersion objective  
  • Numerical Aperture (NA): 1.3  
  • Resolution: 0.21 μm  
• Image Count:  
  • 350 positive images (parasite densities: 1–100 amastigotes per image)  
  • 220 negative images (no parasites, controls)  
  • Total: 570 images 

Dataset 2  

• Patients: Additional cohort (6 patients)
• Samples: Same smear preparation method
• Imaging Setup:  
  • Microscope: Bresser Erudit DLX (portable, battery-powered)  
  • Camera: BRESSER MikroCam SP 5.0  
  • Magnification: 100× oil immersion objective  
  • Numerical Aperture (NA): 1.25  
  • Resolution: 0.22 μm  
• Image Count:  
  • 106 positive images  
  • 58 negative images  
  • Total: 164 images   

👉 Dataset 2 folder = Dataset 1 images + Dataset 2 images (extended dataset). 

3. Directory Structure

Dataset_1/
│
├── Images/
│   ├── train/   # dataset_1_image_1.png ... dataset_1_image_398.png
│   ├── val/     # continues numbering from train
│   └── test/
│
└── Labels/
    ├── train/   # dataset_1_image_1.txt ...
    ├── val/
    └── test/

Dataset_2/
│
├── Images/
│   ├── train/   # contains both dataset_1 and dataset_2 images
│   ├── val/
│   └── test/
│
└── Labels/
    ├── train/
    ├── val/
    └── test/

• Naming Convention: 
  • dataset_1_image_X.png for Dataset 1 images  
  • dataset_2_image_X.png for Dataset 2 additional images  
  • Labels follow the same numbering with .txt extension  
• Splits:
  • Train, validation, and test sets are sequential
  • Example: Train = images 1–398, Val = 399–…, Test = continues onward

4. Labels & Schema  

• Image format: .png
• Label format: .txt (YOLO-style bounding boxes)
• Each line = one object (parasite body)

Format: 

• class_id:  
  • 0 = parasite  
• Coordinates are normalized by image width and height.  
• After class_id, the values are given in pairs: (x1 y1 x2 y2 … x4 y4)
• Every .png has a corresponding .txt file in the same split (train/, val/, test/).

5. Dataset Relations  

• Dataset 1 = base dataset (lab microscope, high-quality).  
• Dataset 2 = superset (Dataset_1 + portable microscope data).  
• Train/Val/Test subsets are disjoint (sequential indexing prevents data leakage).