Colony Detection (ASTPredict)

This repository contains the setup and inference scripts for the Colony Detection. It uses a trained model to detect colonies in images, generating visual annotations and statistical metadata.

Project Structure

• setup.sh: Bash script to install all required dependencies.
• inference.py: Main Python script to run model predictions.
• best.pt: The trained YOLOv8 model weights (must be present in the root directory).
• test_images/: Default directory for placing input images.
• Results/: Output directory (auto-generated) where annotated images and prediction JSONs are saved.

---

1. Environment Setup

Before running the inference script, you need to install the required Python packages (ultralytics for YOLO, and opencv-python).

Open your terminal, navigate to this directory, and run the setup script:

# Make the script executable (macOS/Linux)
chmod +x setup.sh

# Run the setup script
./setup.sh

---

2. Running Inference

The inference.py script allows you to process either a full directory of images or a single image. It will output annotated images with bounding boxes and a predictions.json file containing all the statistical data (bounding box coordinates, confidence scores, etc.).

Basic Usage (Batch Processing)(Use python3)

If you want to run the inference on sample images in ./test_images directory, simply run:

python3 inference.py

Custom Input Directory

To run batch inference on a specific folder of images:

python3 inference.py -i /path/to/your/custom_directory/

Single Image Processing

To run inference on just one specific image:

python3 inference.py -i /path/to/image.jpg

Adjusting Confidence Threshold

By default, the script only considers detections with a confidence score of 0.25 or higher. You can adjust this using the -c or --conf flag. For example, to require a 50% confidence:

python3 inference.py -i /path/to/image.jpg -c 0.5

---

3. Viewing Results

After inference is complete, check the ./Results folder.

• Results/images/: Contains copies of your images drawn with bounding boxes and confidence scores.
• Results/predictions/predictions.json: Contains detailed JSON output of every detection (class, confidence, and normalized/pixel bounding box coordinates).

4. Class - Species Table

Class	Bacteria Species
class0	Actinobacillus equuli
class1	Actinobacillus pleuropneumoniae
class2	Aeromonas hydrophila
class3	Bacillus cereus
class4	Bibersteinia trehalosi
class5	Bordetella bronchiseptica
class6	Brucella ovis
class7	Clostridium perfringens
class8	Corynebacterium pseudotuberculosis
class9	Erysipelothrix rhusiopathiae
class10	Escherichia coli
class11	Glaesserella parasuis
class12	Klebsiella pneumoniae
class13	Listeria monocytogenes
class14	Paenibacillus larvae
class15	Pasteurella multocida
class16	Proteus mirabilis
class17	Pseudomonas aeruginosa
class18	Rhodococcus equi
class19	Salmonella enterica
class20	Staphylococcus aureus
class21	Staphylococcus hyicus
class22	Streptococcus agalactiae
class23	Trueperella pyogenes