# AI use cases for 4DAnyone (ant-research/4DAnyone) Repo: [ant-research/4DAnyone](https://github.com/ant-research/4DAnyone) | Use cases: 3 | Page: https://dedupedai.com/use-cases/ant-research-4danyone/ 4DAnyone is a Python research repo that turns a casual monocular video of one person into consistent multiview videos for downstream 4D Gaussian Splatting reconstruction. This file holds every use case of this repo. Each one starts with a "## " heading and ends with its URL. - Turn a casual clip into a 24-view orbit set: https://dedupedai.com/use-cases/ant-research-4danyone/multiview-video-from-casual-clip/ - Preview camera layouts in a browser GUI: https://dedupedai.com/use-cases/ant-research-4danyone/preview-camera-layouts-in-gui/ - Run a quick 6-view test on a consumer GPU: https://dedupedai.com/use-cases/ant-research-4danyone/quick-six-view-test-on-consumer-gpu/ ## How to turn a casual clip into a 24-view orbit set with 4DAnyone Repo: [ant-research/4DAnyone](https://github.com/ant-research/4DAnyone) | Difficulty: Intermediate | Time: 1 hour | Basis: From the README Generate 24 consistent target-view videos of one person from a single monocular video. **Who it is for:** A researcher or developer preparing multiview training data for 4D Gaussian Splatting when they only have one casual video of a person. **The problem:** Multiview capture of a person needs many synchronized cameras, which most people do not have. **The outcome:** A folder with cameras.json, pose skeleton videos, and 24 target videos covering a full 360 degree orbit. ### Steps 1. `git clone https://github.com/ant-research/4DAnyone.git` 2. `cd 4DAnyone` 3. `git submodule update --init third_party/GVHMR` 4. `conda create -n 4danyone python=3.11 -y` then `conda activate 4danyone` 5. `pip install -r requirements.txt` 6. `python inference.py --video_path "data/source/pexels/2785536-uhd_2160_3840_25fps.mp4" --output_dir "data/fdanyone/pexels/2785536-uhd_2160_3840_25fps" --views_per_layer 24` (inferred, check the README) ### Prompt to paste into your agent ```text In the 4DAnyone repo, replace the example video path with and run inference.py with --views_per_layer 24 writing to . Expected output: cameras.json, skeletons/, and videos/ with 24 target videos. Verify that cameras.json lists 24 views and videos/ contains 24 .mp4 files. ``` ### Caveat The README asks for a 121-frame, 1080p or higher, 9:16 single-person video and notes that inference averages about 27 seconds per 121-frame video on one RTX 4090. Dual-use notice: only use this on software, sites and systems you own or are authorized to test. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/ant-research-4danyone/multiview-video-from-casual-clip/ ## How to preview camera layouts in a browser GUI with 4DAnyone Repo: [ant-research/4DAnyone](https://github.com/ant-research/4DAnyone) | Difficulty: Intermediate | Time: 1 hour | Basis: From the README Run the Gradio GUI to inspect inference results and visualize target views with Rerun. **Who it is for:** A developer or artist who wants to see the generated views and camera layout before committing to a full pipeline run. **The problem:** Reading cameras.json and folders of MP4 files by hand makes it hard to judge whether a camera layout suits the subject. **The outcome:** A local web page at port 7860 showing the inference results for an existing output directory. ### Steps 1. `git clone https://github.com/ant-research/4DAnyone.git` 2. `cd 4DAnyone` 3. `pip install -r requirements.txt` 4. `pip install -r requirements-gui.txt` 5. `python app.py --output_dir "data/fdanyone/pexels/2785536-uhd_2160_3840_25fps" --server_port 7860` (inferred, check the README) 6. Open http://127.0.0.1:7860 in your browser. ### Prompt to paste into your agent ```text In the 4DAnyone repo, run app.py with --output_dir set to and --server_port 7860, then confirm the Gradio page loads at http://127.0.0.1:7860. Expected output: an interactive view of the inference results. Verify that the output directory contains cameras.json and videos/ before launching. ``` ### Caveat The README says the GUI needs requirements-gui.txt installed in the same 4danyone environment, and a remote GPU server needs an SSH port forward to reach the page. Dual-use notice: only use this on software, sites and systems you own or are authorized to test. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/ant-research-4danyone/preview-camera-layouts-in-gui/ ## How to run a quick 6-view test on a consumer GPU with 4DAnyone Repo: [ant-research/4DAnyone](https://github.com/ant-research/4DAnyone) | Difficulty: Intermediate | Time: 1 hour | Basis: From the README Produce a compact 360 degree six-view set with the Turbo default to check the pipeline works. **Who it is for:** A developer with a consumer NVIDIA GPU who wants to confirm the install works before running heavier camera layouts. **The problem:** Full runs with many views take longer, so a first test needs a small, fast check. **The outcome:** Six target videos in a 360 degree orbit, produced with the default distilled Turbo model. ### Steps 1. `git clone https://github.com/ant-research/4DAnyone.git` 2. `cd 4DAnyone` 3. `git submodule update --init third_party/GVHMR` 4. `conda create -n 4danyone python=3.11 -y` then `conda activate 4danyone` 5. `pip install -r requirements.txt` 6. `python scripts/download_example.py` 7. `python inference.py --video_path "data/source/pexels/2785536-uhd_2160_3840_25fps.mp4" --output_dir "data/fdanyone/pexels/2785536-uhd_2160_3840_25fps" --views_per_layer 6` (inferred, check the README) ### Prompt to paste into your agent ```text In the 4DAnyone repo, run scripts/download_example.py and then inference.py with --views_per_layer 6 on the example video, writing to . Expected output: six target videos and cameras.json. Verify that videos/ holds 6 .mp4 files and that metadata.json records the run settings. ``` ### Caveat The README says the example video path is the one used throughout, and the Turbo model is the default, so the Base model is not used unless you change the run settings. Dual-use notice: only use this on software, sites and systems you own or are authorized to test. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/ant-research-4danyone/quick-six-view-test-on-consumer-gpu/