Skip to content

Display interface for manual inspection

The Display module provides an interactive web-based dashboard for visualizing and analyzing image data using Dash and Bootstrap components.

Key Features

  • Interactive File Explorer: Dropdown interface to select images from a dictionary
  • Image Display: Real-time image preview in the dashboard
  • Multi-Detector Support: Run different analysis detectors:
  • TextDetector (text extraction and translation)
  • ColorDetector (color analysis)
  • VQA (Visual Question Answering with image summaries)
  • Dynamic Settings: Context-aware settings panels that appear based on selected detector
  • Real-time Analysis: Execute analysis on selected images with loading indicators
  • Results Visualization: JSON-style table display of analysis results
  • Customizable Questions: Text area for entering custom VQA questions (one per line)
  • Privacy Controls: Environment variable configuration for privacy disclosure acceptance

Usage

from ammico.display import AnalysisExplorer

explorer = AnalysisExplorer(mydict=image_dict)
explorer.run_server(port=8050)

Features

  • TextDetector Settings: Privacy disclosure environment variable configuration
  • ColorDetector Settings: Delta-E method selection dropdown
  • VQA Settings:
  • Analysis type selection (summary, questions, or both)
  • Custom questions textarea (shown/hidden based on analysis type)
  • Model selection (base/large)

Output

Displays analysis results in an interactive table format, showing all extracted features and metadata for the selected image.

Workflow

flowchart TD
    Start([Initialize AnalysisExplorer]) --> RunServer[Explorer.run_server]
    RunServer --> Dashboard[Dash Dashboard Loaded]

    subgraph Dashboard Interaction
        Dashboard --> SelectImg{Select Image}
        SelectImg --> UpdatePic[Update Picture View]

        Dashboard --> SelectDet{Select Detector}
        SelectDet --> UpdateSet[Update Settings Panel]

        UpdateSet --> Settings{Adjust Settings}

        Settings --> ClickRun{Click Run Detector}
        ClickRun --> CheckDet[Check Selected Detector]
    end

    CheckDet -- TextDetector --> TextDet[Init TextDetector]
    CheckDet -- ColorDetector --> ColorDet[Init ColorDetector]
    CheckDet -- VQA --> VQADet[Init ImageSummaryDetector]

    TextDet --> RunMeasure[Run analyse_image]
    ColorDet --> RunMeasure
    VQADet --> RunMeasure

    RunMeasure --> Format[Format Output as Table]
    Format --> UpdateTable[Update JSON Viewer]

    UpdateTable --> Dashboard