Perspective

Platform: CUSTOM | Category: Reversing | Type: Challenge | Difficulty: Medium | OS: N/A | **Author: XESXOR Techniques: 3D point cloud analysis, Float data extraction from binary, Orthographic projection visualization, PE binary analysis

Summary

Walkthrough of the CUSTOM Reversing Perspective. Task: Windows PE64 OpenGL/GLFW 3D application with camera controls and hidden message. Solution: Extract float32 data from binary, convert to 3D points, apply orthographic XY projection to reveal flag

Description

An artist's perspective is always unique

File: perspective.zip contains perspective.exe

Analysis

1. File Identification

 bash read-only • copy allowed Copy
file perspective.exe
# perspective.exe: PE32+ executable (console) x86-64, for MS Windows

This is a Windows PE64 application. When running (or analyzing strings) we can see:

• OpenGL/GLFW 3D application
• Camera controls: WASD (movement), Arrow keys (rotation), R (reset)
• Text: "The Artist's Perspective - What do you see?" ‍​‌‌​​​​‌​‌‌​​‌‌​​‌‌​​‌‌​​‌‌​​‌​​​​‌‌​​​​​‌‌​​‌‌​​‌‌​​‌​‌​‌‌​​‌​‌‍

2. Finding Data

Binary analysis reveals a large block of float32 data at offset 0x5d3c0:

 bash read-only • copy allowed Copy
xxd -s 0x5d3c0 -l 64 perspective.exe
# Float number patterns visible in little-endian format

3. Extracting Points

Found 1366 3D points (4098 float values = 1366 * 3 coordinates):

• X range: -24.871 to 24.294
• Y range: -14.983 to 15.000
• Z range: -9.864 to 24.920

Solution

Extracting and Visualizing Points

 python read-only • copy allowed Copy
#!/usr/bin/env python3
import struct
import numpy as np
import matplotlib.pyplot as plt
‍​‌‌​​​​‌​‌‌​​‌‌​​‌‌​​‌‌​​‌‌​​‌​​​​‌‌​​​​​‌‌​​‌‌​​‌‌​​‌​‌​‌‌​​‌​‌‍

# Read the binary
with open('perspective.exe', 'rb') as f:
    data = f.read()

# Extract float data
offset = 0x5d3c0
num_floats = 4098  # 1366 points * 3 coordinates

floats = []
for i in range(num_floats):
    val = struct.unpack('<f', data[offset + i*4 : offset + i*4 + 4])[0]
    floats.append(val)

# Convert to 3D points
points = np.array([
    (floats[i], floats[i+1], floats[i+2]) 
    for i in range(0, len(floats)-2, 3)
])

print(f"Total points: {len(points)}")
print(f"X range: {points[:,0].min():.3f} to {points[:,0].max():.3f}")
print(f"Y range: {points[:,1].min():.3f} to {points[:,1].max():.3f}")
print(f"Z range: {points[:,2].min():.3f} to {points[:,2].max():.3f}")
‍​‌‌​​​​‌​‌‌​​‌‌​​‌‌​​‌‌​​‌‌​​‌​​​​‌‌​​​​​‌‌​​‌‌​​‌‌​​‌​‌​‌‌​​‌​‌‍

# XY projection visualization (top-down view)
plt.figure(figsize=(20, 10))
plt.scatter(points[:, 0], points[:, 1], s=8, c='black')
plt.axis('equal')
plt.title('XY Projection - Top-Down View')
plt.savefig('flag.png', dpi=150, bbox_inches='tight')
plt.show()

Result

When projecting points onto the XY plane (orthographic top-down projection), the points form readable text:

0ops{REDACTED}

Alternative Approaches

1. Running in Windows/Wine — can try to find the correct camera angle manually
2. 3D visualization — use matplotlib 3D scatter for interactive rotation
3. Different projections — try XZ, YZ projections if XY doesn't yield results ‍​‌‌​​​​‌​‌‌​​‌‌​​‌‌​​‌‌​​‌‌​​‌​​​​‌‌​​​​​‌‌​​‌‌​​‌‌​​‌​‌​‌‌​​‌​‌‍

Theory

The challenge uses a classic steganography technique in 3D space:

• Points are positioned so that at a certain viewing angle they form text
• Orthographic projection (without perspective) best reveals the hidden message
• Leetspeak in the flag: "Perspective is everything" ‍​‌‌​​​​‌​‌‌​​‌‌​​‌‌​​‌‌​​‌‌​​‌​​​​‌‌​​​​​‌‌​​‌‌​​‌‌​​‌​‌​‌‌​​‌​‌‍

Key Takeaways / Lessons

• Mastering 3D point cloud analysis, Float data extraction from binary, Orthographic projection visualization, PE binary analysis was key to this challenge.
• Always check for common misconfigurations and verify assumptions.

---

Crafted by XESXOR Research • 2025-12-20 • CUSTOM • Perspective • Difficulty: Medium