AI Tools
Back to Edit ImagesImage Selection
No selection loaded
Select a selection from the dropdown above to work with specific images
AI Tools Overview
Use these AI-powered tools to enhance your image metadata. Each tool will process only the images from your selected selection. Make sure to select a selection first!
Image Cropping Tool
The foundation of your workflow - crop images to extract text and prepare them for AI processing
What it does:
- Interactive cropping interface for precise text extraction
- Creates cropped versions for AI text extraction
- Essential first step for all AI processing
Workflow:
How the cropping tool works:
This tool provides an interactive interface where you can select and crop specific areas of your images that contain text. The cropped versions are then used by the AI tools below to extract text, detect languages, and generate descriptions. This is the essential first step in your AI-powered workflow.
Distribution Files Check
Verify the existence of EPS, PNG, and SVG files in your distribution folders
What it does:
- Scans 04_Distribution/EPS, PNG, SVG folders
- Updates EPS, PNG, SVG ready flags in database
- No AI required - pure file system check
File Structure:
How the distribution files check works:
This tool automatically scans your file system to check if EPS, PNG, and SVG versions of your images exist in the distribution folders. For each image in your selected search, it looks for corresponding files in the 04_Distribution/EPS, 04_Distribution/PNG, and 04_Distribution/SVG folders. The tool updates the database with the current status of these files, which is reflected in the readiness percentage on the edit images page.
PNG to JPG Conversion
Convert PNG files to JPG format for images that have PNG files ready
What it does:
- Processes images with png_ready=True in database
- Converts PNG files to high-quality JPG format with 100px white padding
- Saves JPG files in 04_Distribution/JPG folder
- Updates jpg_ready flag in database
File Structure:
How the PNG to JPG conversion works:
This tool automatically converts PNG files to JPG format for images that have PNG files ready in the distribution folder. It processes only images where png_ready=True in the database and creates high-quality JPG versions in the 04_Distribution/JPG folder. The tool handles transparency by adding a white background, adds 100 pixels of white padding on all sides for better presentation, and updates the jpg_ready flag in the database after successful conversion.
Image to PDF Conversion
Convert images to PDF format - one image per PDF with professional layout
What it does:
- Processes images with png_ready=True or jpg_ready=True
- Creates one PDF per image with A4 page size
- Centers images on page with professional margins
- Saves PDF files in 04_Distribution/PDF folder
- Updates pdf_ready flag in database
File Structure:
How the image to PDF conversion works:
This tool automatically converts images to PDF format for images that have PNG or JPG files ready in the distribution folder. It processes only images where png_ready=True or jpg_ready=True in the database and creates professional PDF versions in the 04_Distribution/PDF folder. Each image becomes its own PDF with A4 page size, centered layout, and proper margins. The tool prefers PNG files over JPG files when both are available and updates the pdf_ready flag in the database after successful conversion.
AI Service Configuration
Before using AI tools, test your AI service configuration to ensure it's working properly. Note: AI tools require either a local LLM (LM Studio) or OpenAI API to be configured and running.
Important: Text Extraction & Language Detection Requirements
The Text Extraction and Language Detection tools below have different requirements. Workflow: First crop your images, then extract text, then detect the language of the extracted text.
- Text Extraction: Images must have cropped versions available (created using the Image Cropping Tool)
- Language Detection: Images must have caption text already extracted (use Text Extraction first)
Text Extraction
OCRExtract text content from cropped images using AI-powered OCR. This tool processes images that have been cropped and saves the extracted text to the caption field.
<strong>How it works:</strong> Uses AI to read text from cropped images and saves it to the caption field. Only processes images that have been cropped using the cropping tool.
Language Detection
NLPAutomatically detect the language of caption text extracted from images. This helps categorize images by language for better organization.
How it works: Analyzes the caption text content to determine the language and saves it to the caption language field.
Description Generation
AIGenerate detailed descriptions of images using AI vision analysis. Creates comprehensive descriptions based on visual content.
How it works: Uses AI vision analysis to create detailed descriptions of image content, including objects, scenes, and visual elements.
JPG Description Generation
AIGenerate specialized descriptions for JPG files created from PNG conversion. Optimized for black-and-white vectorial illustrations focusing on style and composition.
How it works: Analyzes converted JPG files (from PNG to JPG tool) to create specialized descriptions focused on artistic style, subject matter, and composition for black-and-white vectorial illustrations.
Short JPG Description Generation
AIGenerate short, concise descriptions for JPG files that already have long descriptions. Creates 2-sentence descriptions directly describing the subject without introductory phrases.
How it works: Only processes images that already have long JPG descriptions and creates concise 2-sentence descriptions. The AI is instructed to describe the subject directly without using phrases like "An illustration of..." or "The image shows...". It will not overwrite existing short descriptions.
AI Tags Generation
TagsGenerate relevant tags for images using AI analysis. Creates searchable tags based on visual content and context.
Requirement: This tool will only process images that have existing caption text (extracted from previous OCR processing).
English Text Generation From Caption
TranslationTranslate captions from other languages to English. Converts non-English caption text into English for better searchability.
How it works: Takes the existing caption field and translates it from other languages to English, then saves the translated text to the English text field.
English Text from JPG Description
GenerationGenerate English text from existing JPG descriptions. Creates a concise 5-word English text summary based on the JPG description field.
How it works: Uses the existing JPG description field to generate a concise 5-word English text summary, which is useful for images that have JPG descriptions but need English text for searchability and context.
English Text from Filename
TranslationTranslate original filenames to English. Converts non-English filenames to English or copies English filenames as-is.
How it works: Takes the original filename (without extension) and translates it to English if it's in another language. If the filename is already in English, it copies it as-is. This is useful for images with descriptive filenames that need to be converted to English text.
Create New Filenames
NamingGenerate web-safe snake_case filenames from English text. Creates descriptive, searchable filenames for better organization.
How it works: Takes the English text field and converts it to a web-safe snake_case filename. For example: "Anatomical scientific illustrative" becomes "anatomical_scientific_illustrative".
Metadata Processing
YAMLProcess metadata.yaml files to extract book information and populate database fields. Automatically finds and parses metadata files from the parent directory of "01_Raw_Scans".
How it works: Looks for metadata.yaml files in the parent directory of "01_Raw_Scans" and extracts information like Book ID, Title, Author, Publisher, Year, Region, Subject, Illustrator, and Engraver. Missing or incomplete files are handled gracefully.
Error Checking
ValidationCheck for errors and inconsistencies in AI-generated content. Validates and corrects issues in descriptions, tags, and text fields.
How it works: Analyzes AI-generated content for errors, inconsistencies, and issues, then provides corrections and improvements.