Update

README.md

@@ -5,7 +5,11 @@ colorFrom: blue
 colorTo: green
 sdk: gradio
 sdk_version: 5.0.0
-app_port: 7860
+python_version: "3.10"
+app_file: gradio_app.py
+suggested_hardware: "cpu-basic"
+models:
+  - blanchon/image-gs-models-utils
 pinned: false
 ---
 
@@ -52,64 +56,77 @@ pinned: false
 </div>
 
 ## Setup
+
 1. Create a dedicated Python environment and install the dependencies
-    ```bash
-    git clone https://github.com/NYU-ICL/image-gs.git
-    cd image-gs
-    conda env create -f environment.yml
-    conda activate image-gs
-    pip install git+https://github.com/rahul-goel/fused-ssim/ --no-build-isolation
-    cd gsplat
-    pip install -e ".[dev]"
-    cd ..
-    ```
+   ```bash
+   git clone https://github.com/NYU-ICL/image-gs.git
+   cd image-gs
+   conda env create -f environment.yml
+   conda activate image-gs
+   pip install git+https://github.com/rahul-goel/fused-ssim/ --no-build-isolation
+   cd gsplat
+   pip install -e ".[dev]"
+   cd ..
+   ```
 2. Download the image and texture datasets from [OneDrive](https://1drv.ms/u/c/3a8968df8a027819/EeshjZJlMtdCmvvmESiN2pABM71EDaoLYmEwuOvecg0tAA?e=GybqBv) and organize the folder structure as follows
-    ```
-    image-gs
-    └── media
-        ├── images
-        └── textures
-    ```
+   ```
+   image-gs
+   └── media
+       ├── images
+       └── textures
+   ```
 3. (Optional) To run saliency-guided Gaussian position initialization, download the pre-trained [EML-Net](https://github.com/SenJia/EML-NET-Saliency) models ([res_imagenet.pth](https://drive.google.com/open?id=1-a494canr9qWKLdm-DUDMgbGwtlAJz71), [res_places.pth](https://drive.google.com/open?id=18nRz0JSRICLqnLQtAvq01azZAsH0SEzS), [res_decoder.pth](https://drive.google.com/open?id=1vwrkz3eX-AMtXQE08oivGMwS4lKB74sH)) and place them under the `models/emlnet/` folder
-    ```
-    image-gs
-    └── models
-        └── emlnet
-            ├── res_decoder.pth
-            ├── res_imagenet.pth
-            └── res_places.pth
-    ```
+   ```
+   image-gs
+   └── models
+       └── emlnet
+           ├── res_decoder.pth
+           ├── res_imagenet.pth
+           └── res_places.pth
+   ```
 
 ## Quick Start
 
-#### Image Compression 
+#### Image Compression
+
 - Optimize an Image-GS representation for an input image `anime-1_2k.png` using `10000` Gaussians with half-precision parameters
+
 ```bash
 python main.py --input_path="images/anime-1_2k.png" --exp_name="test/anime-1_2k" --num_gaussians=10000 --quantize
 ```
+
 - Render the corresponding optimized Image-GS representation at a new resolution with height `4000` (aspect ratio is maintained)
+
 ```bash
 python main.py --input_path="images/anime-1_2k.png" --exp_name="test/anime-1_2k" --num_gaussians=10000 --quantize --eval --render_height=4000
 ```
 
 #### Texture Stack Compression
+
 - Optimize an Image-GS representation for an input texture stack `alarm-clock_2k` using `30000` Gaussians with half-precision parameters
+
 ```bash
 python main.py --input_path="textures/alarm-clock_2k" --exp_name="test/alarm-clock_2k" --num_gaussians=30000 --quantize
 ```
+
 - Render the corresponding optimized Image-GS representation at a new resolution with height `3000` (aspect ratio is maintained)
+
 ```bash
 python main.py --input_path="textures/alarm-clock_2k" --exp_name="test/alarm-clock_2k" --num_gaussians=30000 --quantize  --eval --render_height=3000
 ```
 
 #### Control bit precision of Gaussian parameters
+
 - Optimize an Image-GS representation for an input image `anime-1_2k.png` using `10000` Gaussians with 12-bit-precision parameters
+
 ```bash
 python main.py --input_path="images/anime-1_2k.png" --exp_name="test/anime-1_2k" --num_gaussians=10000 --quantize --pos_bits=12 --scale_bits 12 --rot_bits 12 --feat_bits 12
 ```
 
-#### Switch to saliency-guided Gaussian position initialization 
+#### Switch to saliency-guided Gaussian position initialization
+
 - Optimize an Image-GS representation for an input image `anime-1_2k.png` using `10000` Gaussians with half-precision parameters and saliency-guided initialization
+
 ```bash
 python main.py --input_path="images/anime-1_2k.png" --exp_name="test/anime-1_2k" --num_gaussians=10000 --quantize --init_mode="saliency"
 ```
@@ -121,11 +138,13 @@ We provide a user-friendly web interface built with Gradio for easy experimentat
 ### Setup for Web Interface
 
 1. Install Gradio (in addition to the main dependencies):
+
 ```bash
-pip install gradio>=4.0.0
+pip install gradio>=5.0.0
 ```
 
 2. Launch the web interface:
+
 ```bash
 python gradio_app.py
 ```
@@ -145,18 +164,20 @@ The Gradio interface provides:
 
 ### Interface Sections
 
-1. **Configuration Panel**: 
+1. **Configuration Panel**:
+
    - Basic parameters (number of Gaussians, training steps)
    - Quantization settings for memory efficiency
    - Initialization modes (gradient, saliency, random)
    - Advanced optimization parameters (learning rates, loss weights)
 
-2. **Training Progress**: 
+2. **Training Progress**:
+
    - Real-time streaming logs
    - Current render and Gaussian visualization updates
    - Training status and control buttons
 
-3. **Results Display**: 
+3. **Results Display**:
    - Final optimized image
    - Gradient and saliency maps used for initialization
    - Download capabilities for all results
@@ -170,13 +191,16 @@ The Gradio interface provides:
 - For quick tests, reduce **max steps** to 500-1000
 
 ### Command Line Arguments
+
 Please refer to `cfgs/default.yaml` for the full list of arguments and their default values.
 
 **Post-optimization rendering**
+
 - `--eval` render the optimized Image-GS representation.
 - `--render_height` image height for rendering (aspect ratio is maintained).
 
-**Bit precision control**: 32 bits (float32) per dimension by default 
+**Bit precision control**: 32 bits (float32) per dimension by default
+
 - `--quantize` enable bit precision control of Gaussian parameters.
 - `--pos_bits` bit precision of individual coordinate dimension.
 - `--scale_bits` bit precision of individual scale dimension.
@@ -184,18 +208,21 @@ Please refer to `cfgs/default.yaml` for the full list of arguments and their def
 - `--feat_bits` bit precision of individual feature dimension.
 
 **Logging**
+
 - `--exp_name` path to the logging directory.
 - `--vis_gaussians`: visualize Gaussians during optimization.
 - `--save_image_steps` frequency of rendering intermediate results during optimization.
 - `--save_ckpt_steps` frequency of checkpointing during optimization.
 
 **Input image**
+
 - `--input_path` path to an image file or a directory containing a texture stack.
 - `--downsample` load a downsampled version of the input image or texture stack as the optimization target to evaluate image upsampling performance.
 - `--downsample_ratio` downsampling ratio.
 - `--gamma` optimize in a gamma-corrected space, modify with caution.
 
 **Gaussian**
+
 - `--num_gaussians` number of Gaussians (for compression rate control).
 - `--init_scale` initial Gaussian scale in number of pixels.
 - `--disable_topk_norm` disable top-K normalization.
@@ -204,6 +231,7 @@ Please refer to `cfgs/default.yaml` for the full list of arguments and their def
 - `--init_random_ratio` ratio of Gaussians with randomly initialized position.
 
 **Optimization**
+
 - `--disable_tiles` disable tile-based rendering (warning: optimization and rendering without tiles will be way slower).
 - `--max_steps` maximum number of optimization steps.
 - `--pos_lr` Gaussian position learning rate.
@@ -214,13 +242,17 @@ Please refer to `cfgs/default.yaml` for the full list of arguments and their def
 - `--disable_prog_optim` disable error-guided progressive optimization.
 
 ## Acknowledgements
+
 We would like to thank the [gsplat](https://github.com/nerfstudio-project/gsplat) team, and the authors of [3DGS](https://github.com/graphdeco-inria/gaussian-splatting), [fused-ssim](https://github.com/rahul-goel/fused-ssim), and [EML-Net](https://github.com/SenJia/EML-NET-Saliency) for their great work, based on which Image-GS was developed.
 
 ## License
+
 This project is licensed under the terms of the MIT license.
 
 ## Citation
+
 If you find this project helpful to your research, please consider citing [BibTeX](assets/docs/image-gs.bib):
+
 ```bibtex
 @inproceedings{zhang2025image,
   title={Image-gs: Content-adaptive image representation via 2d gaussians},
@@ -229,4 +261,4 @@ If you find this project helpful to your research, please consider citing [BibTe
   pages={1--11},
   year={2025}
 }
-```
+```