Table of Contents

👉 Try the demo: https://github.com/LM-Kit/lm-kit-net-samples/tree/main/console_net/document-intelligence/image-processing/image_normalizer_batch

Image Preprocessing for C# .NET Applications


🎯 Purpose of the Demo

An interactive console app that resizes, rotates, and auto-crops images for vision / OCR pipelines. Pick a single image or a whole folder, then opt in to each operation: thumbnail long-side, rotation (90 / 180 / 270), or auto-crop of uniform borders.

All processing runs on-device.


👥 Industry Target Audience

  • Document capture: normalize phone uploads before OCR.
  • e-commerce / DAM: produce consistent thumbnails for grid views.
  • Vision pipelines: keep input dimensions within model budgets.
  • Compliance: trim metadata-leaking borders before sharing.
  • Archives: bulk-normalize legacy scans into a canonical size and orientation.

🚀 Problem Solved

Every VLM, OCR engine, and image-embedding model has an input budget. Sending a 12-megapixel phone photo straight to the model bloats tokens and hurts accuracy. The demo wraps the three most-used ImageBuffer transforms behind a menu, so the same prompt-driven flow handles one image or a folder of thousands.


💻 Application Overview

Interactive menu (no command-line arguments) with two modes:

Mode What it does
File Process a single image with chosen options. Prompts for thumbnail long-side, rotation, and auto-crop.
Folder Process every supported image in a folder (optionally recursive). Same option prompts.
Quit Exit.

Outputs are written as <basename>_thumb.png, <basename>_rot{deg}.png, <basename>_crop.png in the chosen output directory.

✨ Key Features

  • ImageBuffer.LoadAsRGB(path) and SaveAsPng(path).
  • ImageBuffer.Resize(width, height): preserves aspect ratio when called via long-side prompt.
  • ImageBuffer.Rotate(int degrees): 90, 180, or 270.
  • ImageBuffer.CropAuto(margin): trims uniform borders.
  • Folder mode: per-file error handling, summary at the end.

🧠 Model

  • None. Pure image transforms; no LLM is loaded.

🛠️ Getting Started

📋 Prerequisites

  • .NET 8.0 or later

▶️ Running the Application

git clone https://github.com/LM-Kit/lm-kit-net-samples
cd lm-kit-net-samples/console_net/document-intelligence/image-processing/image_normalizer_batch
dotnet run

Pick a mode from the menu and follow the prompts.

📚 Related Content

Share