Upload 3 files

README.md

@@ -1,3 +1,152 @@
+# **ScenePilot-Bench: A Large-Scale First-Person Dataset and Benchmark for Evaluation of Vision-Language Models in Autonomous Driving**
+<div align="center">
+  <img src="assets/The overall structure.png" width="800px">
+  <p>Figure 1: Overview of the ScenePilot-Bench dataset and evaluation metrics.</p>
+</div>
+
+
+[![Project Page](https://img.shields.io/badge/Project-Website-blue?style=flat-square)](https://github.com/yjwangtj/ScenePilot-Bench)
+[![Dataset](https://img.shields.io/badge/Dataset-Download-green?style=flat-square)](https://huggingface.co/datasets/larswangtj/ScenePilot-4K/tree/main) 
+[![Paper](https://img.shields.io/badge/Paper-Arxiv-red?style=flat-square)](#)
+
+## 📦 Contents Overview
+
+The dataset files in this repository can be grouped into the following categories.
+
 ---
-license: apache-2.0
+
+## 1. Model Weight Files
+
+- **ScenePilot_2.5_3b_200k_merged.zip**  
+- **ScenePilot_2_2b_200k_merged.zip**
+
+These two compressed files contain pretrained model weights obtained by training on a **200k-scale VQA training set** constructed in this work.
+
+- The former corresponds to **Qwen2.5-VL-3B**
+- The latter corresponds to **Qwen2-VL-2B**
+
+Both models are trained using the same dataset and unified training pipeline, and are used in the main experiments and comparison studies.
+
 ---
+
+## 2. Spatial Perception and Annotation Data
+
+- **VGGT.zip**  
+  Contains annotation data related to spatial perception tasks, including:
+  - Ego-vehicle trajectory information
+  - Depth-related information  
+
+  These annotations are used to support experiments involving trajectory prediction and spatial understanding.
+
+- **YOLO.zip**  
+  Provides 2D object detection results for major traffic participants.
+  All detections are generated by a unified detection model and are used as perception inputs for downstream VQA and risk assessment tasks.
+
+- **scene_description.zip**  
+  Contains scene description results generated from the original data, including:
+  - Weather conditions
+  - Road types
+  - Other environmental and semantic attributes  
+
+  These descriptions are used for scene understanding and for constructing balanced dataset splits.
+
+---
+
+## 3. Dataset Split Definition
+
+- **split_train_test_val.zip**
+
+This file contains the **original video-level dataset split**, including:
+- Training set
+- Validation set
+- Test set
+
+All VQA datasets of different scales are constructed **strictly based on this video-level split** to avoid scene-level information leakage.
+
+---
+
+## 4. VQA Datasets
+
+### 4.1 All-VQA
+
+- **All-VQA.zip**
+
+This archive contains all VQA data in JSON format.
+Files are organized according to training, validation, and test splits.
+
+For example:
+- `Deleted_2D_train_vqa_add_new.json`
+- `Deleted_2D_train_vqa_new.json`
+
+These files together form the complete training VQA dataset. Other files correspond to validation and test data.
+
+---
+
+### 4.2 Test-VQA
+
+- **Test-VQA.zip**
+
+This archive contains the **100k-scale VQA test datasets** used in the experiments.
+
+- `Deleted_2D_test_selected_vqa_100k_final.json`  
+  Used as the main test set in the primary experiments.
+
+Additional test sets are provided for generalization studies:
+- Files ending with `europe`, `japan-and-korea`, `us`, and `other` correspond to geographic generalization experiments.
+- Files ending with `left` correspond to left-hand traffic country experiments.
+
+Each test set contains **100k VQA samples**.
+
+---
+
+### 4.3 Train-VQA
+
+- **Train-VQA.zip**
+
+This archive contains training datasets of different scales:
+- **200k VQA**
+- **2000k VQA**
+
+Additional subsets include:
+- Files ending with `china`, used for geographic generalization experiments.
+- Files ending with `right`, used for right-hand traffic country experiments.
+
+---
+
+## 5. Video Index and Download Information
+
+- **video_name_all.xlsx**
+
+This file lists all videos used in the dataset along with their corresponding download links.
+It is provided to support dataset reproduction and access to the original video resources.
+
+---
+
+## 🔧 Data Processing Utility
+
+- **clip.py**
+
+This repository provides a utility script for extracting image frames from raw videos.
+
+The script performs the following operations:
+- Trims a fixed duration from the beginning and end of each video
+- Samples frames at a fixed rate
+- Organizes extracted frames into structured folders
+
+
+## Citation
+
+```bibtex
+@article{scenepilot,
+  title={ScenePilot-Bench: A Large-Scale First-Person Dataset and Benchmark for Evaluation of Vision-Language Models in Autonomous Driving},
+  author={Yujin Wang, Yutong Zheng, Wenxian Fan, Jinlong Hong, Wei Tiana,Haiyang Yu, Bingzhao Gao, Jianqiang Wang, Hong Chen},
+  journal={arXiv preprint},
+  year={2025}
+}
+```
+
+## License
+
+[![License](https://img.shields.io/badge/License-Apache_2.0-blue.svg)](https://opensource.org/licenses/Apache-2.0)
+
+This project is licensed under the Apache License 2.0 - see the [LICENSE](LICENSE) file for details.

clip.py

@@ -0,0 +1,132 @@
+import os
+import cv2
+from tqdm import tqdm
+
+# ================= User Configuration =================
+INPUT_FOLDER = "bus"          # Name of the input video folder
+PREFIX = "08"                 # Output prefix (must match category index)
+
+# Category index reference:
+# 01  street_ca
+# 02  street_au
+# 03  street_cn
+# 04  street_eu
+# 05  street_kr
+# 06  street_us
+# 07  highway
+# 08  bus
+
+ROOT_DIR = r"your_root_path"        # Root directory of raw videos
+CLIPS_DIR = r"your_clips_path" # Output directory for extracted frames
+
+SAMPLING_FPS = 2        # Frames sampled per second
+TRIM_HEAD_SEC = 180     # Trim first N seconds
+TRIM_TAIL_SEC = 180     # Trim last N seconds
+# ======================================================
+
+
+def sanitize_filename(name: str) -> str:
+    """Remove illegal characters from file names."""
+    return "".join(c for c in name if c.isalnum() or c in (" ", "-", "_")).strip()
+
+
+def process_videos(input_dir, clips_dir, prefix, folder_name):
+    output_main_dir = os.path.join(clips_dir, f"{prefix}_{folder_name}")
+    os.makedirs(output_main_dir, exist_ok=True)
+
+    video_files = sorted([
+        f for f in os.listdir(input_dir)
+        if f.lower().endswith(('.mp4', '.mkv', '.avi', '.mov'))
+    ])
+
+    print(f"\nFound {len(video_files)} video files")
+
+    video_index = 0
+
+    # Optional: specify a video file name to start processing from
+    start_from = None  
+
+    for filename in video_files:
+        if start_from is not None:
+            if filename != start_from:
+                print(f"Skip: {filename}")
+                video_index += 1
+                continue
+            else:
+                print(f"Start from specified video: {filename}")
+                start_from = None
+
+        file_path = os.path.join(input_dir, filename)
+        print(f"\nProcessing video: {file_path}")
+
+        cap = cv2.VideoCapture(file_path)
+        if not cap.isOpened():
+            print("Failed to open video file")
+            continue
+
+        fps = cap.get(cv2.CAP_PROP_FPS)
+        total_frames = int(cap.get(cv2.CAP_PROP_FRAME_COUNT))
+        duration = total_frames / fps
+
+        print(f"Duration: {duration:.2f} seconds")
+
+        if duration < TRIM_HEAD_SEC + TRIM_TAIL_SEC:
+            print("Skip: video too short after trimming")
+            cap.release()
+            continue
+
+        start_frame = int(TRIM_HEAD_SEC * fps)
+        end_frame = int(total_frames - TRIM_TAIL_SEC * fps)
+        extract_interval = max(1, int(fps / SAMPLING_FPS))
+
+        video_name = os.path.splitext(filename)[0]
+        safe_video_name = sanitize_filename(video_name)
+        output_subdir = os.path.join(output_main_dir, f"{prefix}_{video_index:05d}")
+        os.makedirs(output_subdir, exist_ok=True)
+
+        cap.set(cv2.CAP_PROP_POS_FRAMES, start_frame)
+        frame_count = 0
+        saved_count = 0
+
+        for _ in tqdm(
+            range(end_frame - start_frame),
+            desc=f"Extracting {safe_video_name}",
+            ncols=100
+        ):
+            ret, frame = cap.read()
+            if not ret:
+                break
+
+            if frame_count % extract_interval == 0:
+                output_file = os.path.join(
+                    output_subdir,
+                    f"{prefix}_{video_index:05d}_{saved_count:06d}.jpg"
+                )
+                cv2.imwrite(
+                    output_file,
+                    frame,
+                    [int(cv2.IMWRITE_JPEG_QUALITY), 90]
+                )
+                saved_count += 1
+
+            frame_count += 1
+
+        cap.release()
+        print(f"Saved {saved_count} frames to {output_subdir}")
+        video_index += 1
+
+
+def main():
+    input_dir = os.path.join(ROOT_DIR, INPUT_FOLDER)
+
+    if not os.path.isdir(input_dir):
+        print(f"Invalid input directory: {input_dir}")
+        return
+
+    print(f"Start processing folder: {input_dir}")
+    process_videos(input_dir, CLIPS_DIR, PREFIX, INPUT_FOLDER)
+    print("\nAll videos processed.")
+
+
+if __name__ == "__main__":
+    main()