Oliverdsfdsf/comic-panels-text-detect

YOLO26n-seg for Comic Panels and Text Detection

This model is a fine-tuned version of YOLO26n-seg specifically designed for detecting and segmenting Panels and Text Bubbles in Comics, Manga, and Manhwa.

🚀 Usage in ebookcc

This model powers ebookcc, an automated tool for comic translation and layout analysis.

Predict

📊 Model Details

• Task: Instance Segmentation
• Classes:
  • Panel: Comic frame borders.
  • Text: Speech bubbles and on-page text.
• Input Size: 1280px (optimized for high-res scans).

🛠 How to use (Ultralytics)

from ultralytics import YOLO

# Load the model
model = YOLO('comic-panels-and-text-detect.pt')

# Predict
results = model.predict(source='comic_page.jpg', conf=0.25, imgsz=1280)

# Show results
results[0].show()