## Making Medical Imaging Fun with Python Reflex


[TOC]

Ideally, whenever I open my Mac, I can launch these open-source Python Reflex apps and, through the Internet, create a community where real people and AI interact together.

Recently, I’ve been building two open-source projects for working with DICOM medical images using Python Reflex:

### 1) `dicom_data_explorer`

With `dicom_data_explorer`, you can browse various public medical imaging datasets and download them through a simple UI.

### 2) `dicom_viewer`

With `dicom_viewer`, you can open and view the downloaded medical images.

This setup is mainly designed for Mac users to avoid common security/permission friction. By default:

* `dicom_data_explorer` downloads medical images into:
`/Users/Shared/DICOM`
* `dicom_viewer` reads medical images from that same directory.

Both projects are written in pure Python (using Reflex) and fully open source. Since I develop on macOS, these projects are highly compatible with Mac environments.

---

## Current Features

* **`dicom_viewer` includes window preset guidance**
It provides explanations of common window presets, so people without a medical background can still “feel” the differences when adjusting image display. Anyone interested in medical imaging can explore and learn through hands-on interaction.

* **`dicom_data_explorer` makes public dataset downloads easy**
With a lightweight UI workflow, you can quickly download publicly available medical imaging data.

---

## Future Expansion Ideas

### A) Add more data sources to `dicom_data_explorer`

* Expand beyond public datasets by supporting multiple PACS sources.
* PACS (Picture Archiving and Communication System) integration could be simplified into a “add data source” action—so users can connect and unify different PACS systems through one interface.

### B) Add value-added extensions to `dicom_viewer`

* Introduce an extension mechanism, such as:

* Calling MATLAB functions locally on macOS
* Sending images out to run predictions via in-house or third-party AI models

### C) Enable Human + AI collaboration inside `dicom_viewer`

Python Reflex makes it easy to build collaborative software and integrate LLMs. In the future, medical imaging research should become much easier for cross-border collaboration—people can discuss and analyze together, and ask AI to assist anytime.

Ultimately, as long as I have a Mac, I can host and run everything locally using Python Reflex applications.

---

## My Local “All-in-One” Collaboration Stack (on Mac)

* **`relack`** (an open-source Slack alternative) for text-based meetings and discussions
* **`flex_livekit_audio_chat`** (a Clubhouse alternative) for real-time voice meetings with anyone worldwide
* **`codoc_in_md`** for collaborative meeting notes (with AI-assisted writing).
It already supports all formats supported by HackMD, making it very user-friendly.
* **`codoc_in_vecdraw`** (a collaborative Google Drawings-style tool) when we need an online whiteboard

---

## How It All Connects

With these tools in place, I can:

1. Use `dicom_data_explorer` to download medical imaging datasets that are accessible
2. Use `dicom_viewer` to open and collaboratively explore the images with others
3. Use chat, voice, notes, and whiteboard tools (plus AI) to discuss and learn together

All project links are in my comment section. After about one month of development, individual features are being completed steadily. The next step is to plan the integration properly—everything remains open source and runs on macOS as Python Reflex applications.

If more people can learn about medical images and bring that understanding to broader audiences, it can help everyone gain deeper insight into health and the human body.

## Open Source Ref:

https://github.com/milochen0418/dicom_viewer
https://github.com/milochen0418/dicom_data_explorer
https://github.com/milochen0418/relack
https://github.com/milochen0418/reflex_livekit_audio_chat
https://github.com/milochen0418/codoc_in_md
https://github.com/milochen0418/codoc_in_vecdraw

^{I’ve been using Python Reflex to develop all of these open-source projects. I’ve been working on them for about a month.}

### dicom_viewer
DICOM Viewer is a lightweight image viewer designed for browsing local DICOM folders. It lets you scan and select a DICOM directory, quickly load images into the viewer, and apply common presets to adjust visualization for easier inspection and comparison.

This project is designed to be beginner-friendly for people new to DICOM, and this README includes a complete, step-by-step usage guide (including how to use presets).

[More...](https://github.com/milochen0418/dicom_viewer/)


### dicom_data_explorer
This app helps you explore public DICOM datasets, inspect series metadata, and download image files to your local machine.

[More...](https://github.com/milochen0418/dicom_data_explorer)


### relack
Relack is a secure, self-hosted team communication platform built with Reflex and Python. It is designed to be a privacy-focused alternative to SaaS solutions like Slack.

[More...](https://github.com/milochen0418/relack)


### reflex_livekit_audio_chat
A real-time audio conferencing application built with Reflex and LiveKit.

[More...](https://github.com/milochen0418/reflex_livekit_audio_chat)


### codoc_in_md
CoDoc in MD is a real-time collaborative Markdown editor built with Reflex. It combines a Monaco-powered writing experience with a live preview panel—so you can write, share, and review Markdown docs in one place.

[More...](https://github.com/milochen0418/codoc_in_md)

### codoc_in_vecdraw
A collaborative vector editor similar to Google Drawings, built with Reflex.

[More...](https://github.com/milochen0418/codoc_in_vecdraw)