ImageAnalysisNodeEditor Documentation

Overview

The ImageAnalysisNodeEditor is a powerful component within the Vantage analytics and data platform that enables users to perform image analysis using various AI methods. It provides options to analyze images using descriptions, captions, Optical Character Recognition (OCR), image classification, or custom prompts. The results are stored in a designated output column for further analysis and processing.

Purpose

The primary purpose of the ImageAnalysisNodeEditor is to connect image data sources with advanced AI-driven analysis features. It works with publicly accessible image URLs and integrates with various AI models, including GPT-4o, Claude 3+, and Gemini models.

Data Expectations

The ImageAnalysisNodeEditor expects the following data:

• Image URLs: Must be contained in a specified column that holds accessible image links.
• Configuration Properties: These properties dictate how the image analysis will operate, including the type of analysis and the column where the results will be stored.

Settings

1. Image Column

• Input Type: Dropdown
• Description: Specifies the column containing the image URLs to be analyzed. The images must be publicly accessible.
• Default Value: image_url
• Behavior: Changing this setting selects a different column from upstream data sources. Selecting an invalid or non-existent column will cause the analysis to fail.

2. Analysis Type

• Input Type: Dropdown
• Description: Determines the method of analysis for the images. Options include:
  • Describe Image: Provides a detailed description of the image contents.
  • Caption: Generates a concise one-sentence caption.
  • Extract Text (OCR): Extracts and returns all visible text from the image.
  • Classify: Categorizes the image type (e.g., photo, chart, diagram).
  • Custom Prompt: Allows users to define a personalized analysis prompt.
• Default Value: describe
• Behavior: Changing the analysis type modifies the processing method of the images, thus affecting the nature of the output generated.

3. Custom Prompt (Visible for Custom Type)

• Input Type: Textarea
• Description: When the custom analysis type is selected, this field allows users to input a specific prompt that accompanies each image for analysis.
• Default Value: "" (empty string)
• Behavior: Any text entered here will dictate the specific request made to the AI, thus shaping the output. Users can also insert dynamic references to upstream columns.

4. Output Column

• Input Type: Text
• Description: The name of the column where the AI’s analysis results will be stored.
• Default Value: ai_vision
• Behavior: Changing this affects where the output will be saved. If a non-unique name is provided, it may overwrite existing outputs.

5. Batch Size

• Input Type: Numeric
• Description: This setting controls the number of images processed in parallel per batch. It directly influences processing speed and resource usage.
• Default Value: 5
• Behavior: A higher batch size may speed up the processing but could lead to increased API quota usage. The range is limited from 1 to 10.

AI Integrations

The ImageAnalysisNodeEditor integrates with advanced AI models to enhance image processing capabilities. It utilizes technology from GPT-4o, Claude 3+, and Gemini models, aiming to provide accurate and nuanced image analysis results.

Billing Impact

Usage of the ImageAnalysisNodeEditor can affect billing depending on:

• The number of images processed.
• The selected analysis type (some types may require more computational resources).
• The batch size, as processing more images concurrently may lead to higher API consumption.

Use Cases & Examples

Use Case 1: E-commerce Product Analysis

A retail company wants to automate product image analysis on their website to enhance product descriptions and improve SEO. They can use the ImageAnalysisNodeEditor to configure it to generate descriptions and captions for product images uploaded to their platform.

Use Case 2: Document Management

A law firm intends to digitize and extract vital information from scanned documents (such as contracts and invoices). By configuring the ImageAnalysisNodeEditor with the OCR option, they can extract text and automate the data entry process.

Concrete Example Configuration

Scenario: Automating E-commerce Product Descriptions

• Image Column: product_image_url
• Analysis Type: describe
• Output Column: product_description
• Batch Size: 5

The setup would allow the firm to quickly generate detailed descriptions for multiple product images simultaneously. The configuration can be expressed in sample data as follows:

{
  "imageColumn": "product_image_url",
  "analysisType": "describe",
  "outputColumn": "product_description",
  "batchSize": 5
}

This configuration would enable the e-commerce site to leverage AI in enhancing product information efficiently and effectively, streamlining content generation and improving customer engagement.