modify readme

Author: Woojin52

README.md

@@ -1,20 +1,48 @@
 <div align="center">
+<h2>
+[CVPR 2025] ABBSPO: Adaptive Bounding Box Scaling and Symmetric Prior Based Orientation Prediction for Detecting Aerial Image Objects
+</h2>
+
+<div>
+    <b>Woojin Lee*</b>&nbsp;&nbsp;&nbsp;
+    <b>Hyugjae Chang*</b>&nbsp;&nbsp;&nbsp;
+    Jaeho Moon&nbsp;&nbsp;&nbsp;
+    Jaehyup Lee<sup>†</sup>&nbsp;&nbsp;&nbsp;
+    Munchurl Kim<sup>†</sup>
+</div>
 
-  <img src="assets/abbspo_logo.png" width="450"/>
-
-  <div>&nbsp;</div>
+<br>
+<div>
+    <sup>*</sup>Co-first authors &nbsp;&nbsp;
+    <sup>†</sup>Co-corresponding authors
+</div>
 
-  <div align="center">
-    <b><font size="5">[CVPR 2025] ABBSPO: Adaptive Bounding Box Scaling and Symmetric Prior Based Orientation Prediction for Detecting Aerial Image Objects</font></b>
-  </div>
+<div>
+    KAIST (Korea Advanced Institute of Science and Technology), KNU
+</div>
 
-  <div>&nbsp;</div>
+<br>
+
+<div>
+    <h4 align="center">
+        <a href="https://kaist-viclab.github.io/ABBSPO_site/" target="_blank">
+            <img src="https://img.shields.io/badge/🏠-Project%20Page-blue">
+        </a>
+        <a href="https://openaccess.thecvf.com/content/CVPR2025/papers/Lee_ABBSPO_Adaptive_Bounding_Box_Scaling_and_Symmetric_Prior_based_Orientation_CVPR_2025_paper.pdf" target="_blank">
+            <img src="https://img.shields.io/badge/Paper-CVPR%202025-b31b1b.svg">
+        </a>
+        <img alt="GitHub Repo stars" src="https://img.shields.io/github/stars/KAIST-VICLab/ABBSPO">
+    </h4>
+</div>
+</div>
 
-  [📄 Paper](https://openaccess.thecvf.com/content/CVPR2025/papers/Lee_ABBSPO_Adaptive_Bounding_Box_Scaling_and_Symmetric_Prior_based_Orientation_CVPR_2025_paper.pdf) |
-  [🌐 Project Page](https://kaist-viclab.github.io/ABBSPO_site/) |
-  [📦 Checkpoints (TBA)](#-pretrained-models)
+---
 
-</div>
+<h4>
+This repository is the official PyTorch implementation of 
+<b>"ABBSPO: Adaptive Bounding Box Scaling and Symmetric Prior Based Orientation Prediction for Detecting Aerial Image Objects"</b>.
+ABBSPO is a weakly supervised oriented object detection (WS-OOD) framework that learns accurate rotated bounding boxes using only horizontal bounding box annotations.
+</h4>
 
 ---
 
@@ -23,24 +51,28 @@
 </p>
 
 <p align="center">
-ABBSPO is a weakly supervised oriented object detection (WS-OOD) framework built upon H2RBox-v2.
-It learns oriented bounding boxes (RBoxes) using only horizontal bounding boxes (HBoxes) via
-<b>Adaptive Bounding Box Scaling (ABBS)</b> and a <b>Symmetric Prior Angle (SPA)</b> loss.
+  <a href="assets/figure2.pdf">[Download high-resolution PDF]</a>
 </p>
 
 ---
 
-# 🗞️ News
+## 📧 News
+- **Mar 2025** — Initial code release  
+- **Jan 2025** — Paper accepted to **CVPR 2025**
+
+---
 
-- **Feb 2025** — Initial code release.  
-- **Jan 2025** — Paper accepted to **CVPR 2025**.  
-- **Coming soon** — Checkpoints for DIOR-R and DOTA-v1.0.
+## 🔧 Tested Environment
+- OS: Ubuntu 20.04
+- Python: 3.8
+- PyTorch: 1.13+
+- CUDA: 11.3 / 11.7
+- GPU: NVIDIA RTX series
 
 ---
 
-# ⚙️ Installation
+## ⚙️ Environment Setup
 
-## Option A — Install using pip & requirements.txt
 ```bash
 git clone https://github.com/KAIST-VICLab/ABBSPO.git
 cd ABBSPO