update mertic

README.md

@@ -61,7 +61,200 @@ Real5-OmniDocBench/
 
 To illustrate the impact of real-world visual distortions on document parsing, we evaluated several mainstream models across all five scenarios in Real5-OmniDocBench. The same metrics as OmniDocBench are used: **Overall↑**, **TextEdit↓**, **FormulaCDM↑**, **TableTEDS↑**, and **Reading OrderEdit↓**.
 
-### 1. Scanning
+### 1. Overall
+
+<table border="1" style="width:100%; border-collapse: collapse;">
+  <tr>
+    <th><strong>Model Type</strong></th>
+    <th><strong>Methods</strong></th>
+    <th><strong>Parameters</strong></th>
+    <th><strong>Overall&#x2191</strong></th>
+    <th><strong>Scanning&#x2191</strong></th>
+    <th><strong>Warping&#x2191</strong></th>
+    <th><strong>Screen-Photography,&#x2191</strong></th>
+    <th><strong>Illumination&#x2191</strong></th>
+    <th><strong>Skew&#x2191</strong></th>
+  </tr>
+  <tr>
+    <td rowspan="2">Pipeline Tools</td>
+    <td>Marker-1.8.2</td>
+    <td>-</td>
+    <td>60.10</td>
+    <td>70.27</td>
+    <td>58.98</td>
+    <td>63.65</td>
+    <td>66.31</td>
+    <td>41.27</td>
+  </tr>
+  <tr>
+    <td>PP-StructureV3</td>
+    <td>-</td>
+    <td>64.45</td>
+    <td>84.68</td>
+    <td>59.34</td>
+    <td>66.89</td>
+    <td>73.38</td>
+    <td>37.98</td>
+  </tr>
+  <tr>
+    <td rowspan="5">General VLMs</td>
+    <td>GPT-5.2</td>
+    <td>-</td>
+    <td>78.66</td>
+    <td>84.43</td>
+    <td>76.26</td>
+    <td>76.75</td>
+    <td>80.88</td>
+    <td>75.00</td>
+  </tr>
+  <tr>
+    <td>Qwen2.5-VL-72B</td>
+    <td>72B</td>
+    <td>86.92</td>
+    <td>86.19</td>
+    <td>87.77</td>
+    <td>86.48</td>
+    <td>87.25</td>
+    <td>86.90</td>
+  </tr>
+  <tr>
+    <td>Gemini-2.5 Pro</td>
+    <td>-</td>
+    <td>88.21</td>
+    <td>89.25</td>
+    <td>87.63</td>
+    <td>87.11</td>
+    <td>87.97</td>
+    <td>89.07</td>
+  </tr>
+  <tr>
+    <td>Qwen3-VL-235B</td>
+    <td>235B</td>
+    <td>88.90</td>
+    <td>89.43</td>
+    <td>89.99</td>
+    <td>89.27</td>
+    <td>89.27</td>
+    <td>86.56</td>
+  </tr>
+  <tr>
+    <td>Gemini-3 Pro</td>
+    <td>-</td>
+    <td>89.24</td>
+    <td>89.47</td>
+    <td>88.90</td>
+    <td>88.86</td>
+    <td>89.53</td>
+    <td><u>89.45</u></td>
+  </tr>
+  <tr>
+    <td rowspan="10">Specialized VLMs</td>
+    <td>Deepseek-OCR 2</td>
+    <td>3B</td>
+    <td>73.01</td>
+    <td>89.59</td>
+    <td>66.53</td>
+    <td>71.65</td>
+    <td>76.02</td>
+    <td>61.28</td>
+  </tr>
+  <tr>
+    <td>Deepseek-OCR</td>
+    <td>3B</td>
+    <td>73.99</td>
+    <td>86.17</td>
+    <td>67.20</td>
+    <td>75.31</td>
+    <td>78.10</td>
+    <td>63.01</td>
+  </tr>
+  <tr>
+    <td>MinerU2-VLM</td>
+    <td>0.9B</td>
+    <td>76.95</td>
+    <td>83.60</td>
+    <td>73.73</td>
+    <td>78.77</td>
+    <td>80.51</td>
+    <td>68.16</td>
+  </tr>
+  <tr>
+    <td>MonkeyOCR-pro-3B</td>
+    <td>3.7B</td>
+    <td>79.49</td>
+    <td>86.94</td>
+    <td>78.90</td>
+    <td>82.44</td>
+    <td>84.71</td>
+    <td>64.47</td>
+  </tr>
+  <tr>
+    <td>Nanonets-OCR-s</td>
+    <td>3B</td>
+    <td>84.19</td>
+    <td>85.52</td>
+    <td>83.56</td>
+    <td>84.86</td>
+    <td>85.01</td>
+    <td>81.98</td>
+  </tr>
+  <tr>
+    <td>dots.ocr</td>
+    <td>3B</td>
+    <td>86.38</td>
+    <td>86.87</td>
+    <td>86.01</td>
+    <td>87.18</td>
+    <td>87.57</td>
+    <td>84.27</td>
+  </tr>
+  <tr>
+    <td>PaddleOCR-VL</td>
+    <td>0.9B</td>
+    <td>85.54</td>
+    <td>92.11</td>
+    <td>85.97</td>
+    <td>82.54</td>
+    <td>89.61</td>
+    <td>77.47</td>
+  </tr>
+  <tr>
+    <td>MinerU2.5</td>
+    <td>1.2B</td>
+    <td>85.61</td>
+    <td>90.06</td>
+    <td>83.76</td>
+    <td>89.41</td>
+    <td>89.57</td>
+    <td>75.24</td>
+  </tr>
+  <tr>
+    <td>GLM-OCR</td>
+    <td>0.9B</td>
+    <td><u>90.32</u></td>
+    <td><u>92.67</u></td>
+    <td><u>90.68</u></td>
+    <td><u>91.75</u></td>
+    <td><u>91.12</u></td>
+    <td>85.39</td>
+  </tr>
+  <tr>
+    <td><strong>PaddleOCR-VL-1.5</strong></td>
+    <td>0.9B</td>
+    <td><strong>92.05</strong></td>
+    <td><strong>93.43</strong></td>
+    <td><strong>91.25</strong></td>
+    <td><strong>91.76</strong></td>
+    <td><strong>92.16</strong></td>
+    <td><strong>91.66</strong></td>
+  </tr>
+</table>
+
+*Overall is the average across five scenarios
+
+---
+
+### 2. Scanning
 
 <table border="1" style="width:100%; border-collapse: collapse;">
   <tr>
@@ -140,7 +333,7 @@ To illustrate the impact of real-world visual distortions on document parsing, w
     <td>0.078</td>
   </tr>
   <tr>
-    <td rowspan="12">Specialized VLMs</td>
+    <td rowspan="14">Specialized VLMs</td>
     <td>Dolphin</td>
     <td>322M</td>
     <td>72.16</td>
@@ -221,6 +414,15 @@ To illustrate the impact of real-world visual distortions on document parsing, w
     <td>84.86</td>
     <td>0.141</td>
   </tr>
+  <tr>
+    <td>Deepseek-OCR 2</td>
+    <td>3B</td>
+    <td>89.59</td>
+    <td>0.055</td>
+    <td>88.55</td>
+    <td>85.72</td>
+    <td>0.056</td>
+  </tr>
   <tr>
     <td>MinerU2.5</td>
     <td>1.2B</td>
@@ -233,19 +435,28 @@ To illustrate the impact of real-world visual distortions on document parsing, w
   <tr>
     <td>PaddleOCR-VL</td>
     <td>0.9B</td>
-    <td><u>92.11</u></td>
+    <td>92.11</td>
     <td><u>0.039</u></td>
-    <td><u>90.35</u></td>
-    <td><u>89.90</u></td>
+    <td>90.35</td>
+    <td>89.90</td>
     <td><u>0.048</u></td>
   </tr>
+  <tr>
+    <td>GLM-OCR</td>
+    <td>0.9B</td>
+    <td><u>92.67</u></td>
+    <td>0.054</td>
+    <td><u>91.10</u></td>
+    <td><strong>92.28</strong></td>
+    <td>0.061</td>
+  </tr>
   <tr>
     <td><strong>PaddleOCR-VL-1.5</strong></td>
     <td>0.9B</td>
     <td><strong>93.43</strong></td>
     <td><strong>0.037</strong></td>
     <td><strong>93.04</strong></td>
-    <td><strong>90.97</strong></td>
+    <td><u>90.97</u></td>
     <td><strong>0.045</strong></td>
   </tr>
 </table>
@@ -254,7 +465,7 @@ To illustrate the impact of real-world visual distortions on document parsing, w
 
 ---
 
-### 2. Warping
+### 3. Warping
 
 <table border="1" style="width:100%; border-collapse: collapse;">
   <tr>
@@ -320,20 +531,20 @@ To illustrate the impact of real-world visual distortions on document parsing, w
     <td>88.90</td>
     <td>0.086</td>
     <td>88.10</td>
-    <td><u>87.20</u></td>
+    <td>87.20</td>
     <td>0.087</td>
   </tr>
   <tr>
     <td>Qwen3-VL-235B-A22B-Instruct</td>
     <td>235B</td>
-    <td><u>89.99</u></td>
+    <td>89.99</td>
     <td><strong>0.051</strong></td>
-    <td><u>89.06</u></td>
+    <td>89.06</td>
     <td>85.95</td>
     <td><u>0.064</u></td>
   </tr>
   <tr>
-    <td rowspan="12">Specialized VLMs</td>
+    <td rowspan="14">Specialized VLMs</td>
     <td>Dolphin-1.5 </td>
     <td>0.3B</td>
     <td>50.50</td>
@@ -351,6 +562,15 @@ To illustrate the impact of real-world visual distortions on document parsing, w
     <td>51.58</td>
     <td>0.247</td>
   </tr>
+  <tr>
+    <td>Deepseek-OCR 2</td>
+    <td>3B</td>
+    <td>66.53</td>
+    <td>0.293</td>
+    <td>70.42</td>
+    <td>58.44</td>
+    <td>0.209</td>
+  </tr>
   <tr>
     <td>Deepseek-OCR</td>
     <td>3B</td>
@@ -432,13 +652,22 @@ To illustrate the impact of real-world visual distortions on document parsing, w
     <td>81.74</td>
     <td>0.093</td>
   </tr>
+  <tr>
+    <td>GLM-OCR</td>
+    <td>0.9B</td>
+    <td><u>90.68</u></td>
+    <td>0.071</td>
+    <td><u>90.30</u></td>
+    <td><strong>88.78</strong></td>
+    <td>0.100</td>
+  </tr>
   <tr>
     <td><strong>PaddleOCR-VL-1.5</strong></td>
     <td>0.9B</td>
     <td><strong>91.25</strong></td>
     <td><u>0.053</u></td>
     <td><strong>90.94</strong></td>
-    <td><strong>88.10</strong></td>
+    <td><u>88.10</u></td>
     <td><strong>0.063</strong></td>
   </tr>
 </table>
@@ -447,7 +676,7 @@ To illustrate the impact of real-world visual distortions on document parsing, w
 
 ---
 
-### 3. Screen-Photography
+### 4. Screen-Photography
 
 <table border="1" style="width:100%; border-collapse: collapse;">
   <tr>
@@ -513,20 +742,20 @@ To illustrate the impact of real-world visual distortions on document parsing, w
     <td>88.86</td>
     <td>0.084</td>
     <td>87.33</td>
-    <td><u>87.65</u></td>
+    <td>87.65</td>
     <td>0.087</td>
   </tr>
   <tr>
     <td>Qwen3-VL-235B-A22B-Instruct</td>
     <td>235B</td>
-    <td><u>89.27</u></td>
-    <td><u>0.068</u></td>
-    <td><u>88.72</u></td>
+    <td>89.27</td>
+    <td>0.068</td>
+    <td>88.72</td>
     <td>85.85</td>
     <td>0.071</td>
   </tr>
   <tr>
-    <td rowspan="12">Specialized VLMs</td>
+    <td rowspan="14">Specialized VLMs</td>
     <td>Dolphin</td>
     <td>322M</td>
     <td>64.29</td>
@@ -544,6 +773,15 @@ To illustrate the impact of real-world visual distortions on document parsing, w
     <td>68.00</td>
     <td>0.177</td>
   </tr>
+  <tr>
+    <td>Deepseek-OCR 2</td>
+    <td>3B</td>
+    <td>71.65</td>
+    <td>0.201</td>
+    <td>73.49</td>
+    <td>61.54</td>
+    <td>0.157</td>
+  </tr>
   <tr>
     <td>Deepseek-OCR</td>
     <td>3B</td>
@@ -620,23 +858,32 @@ To illustrate the impact of real-world visual distortions on document parsing, w
     <td>MinerU2.5</td>
     <td>1.2B</td>
     <td>89.41</td>
-    <td>0.062</td>
+    <td><u>0.062</u></td>
     <td>87.55</td>
     <td>86.83</td>
     <td><strong>0.053</strong></td>
   </tr>
+  <tr>
+    <td>GLM-OCR</td>
+    <td>0.9B</td>
+    <td><u>91.75</u></td>
+    <td>0.063</td>
+    <td><u>89.83</u></td>
+    <td><strong>91.66</strong></td>
+    <td>0.070</td>
+  </tr>
   <tr>
     <td><strong>PaddleOCR-VL-1.5</strong></td>
     <td>0.9B</td>
     <td><strong>91.76</strong></td>
     <td><strong>0.050</strong></td>
     <td><strong>90.88</strong></td>
-    <td><strong>89.38</strong></td>
+    <td><u>89.38</u></td>
     <td><u>0.059</u></td>
   </tr>
 </table>
 
-### 4. Illumination
+### 5. Illumination
 
 <table border="1" style="width:100%; border-collapse: collapse;">
   <tr>
@@ -715,7 +962,7 @@ To illustrate the impact of real-world visual distortions on document parsing, w
     <td>0.080</td>
   </tr>
   <tr>
-    <td rowspan="12">Specialized VLMs</td>
+    <td rowspan="14">Specialized VLMs</td>
     <td>Dolphin</td>
     <td>322M</td>
     <td>67.29</td>
@@ -733,6 +980,15 @@ To illustrate the impact of real-world visual distortions on document parsing, w
     <td>72.69</td>
     <td>0.133</td>
   </tr>
+  <tr>
+    <td>Deepseek-OCR 2</td>
+    <td>3B</td>
+    <td>76.02</td>
+    <td>0.168</td>
+    <td>77.83</td>
+    <td>67.01</td>
+    <td>0.122</td>
+  </tr>
   <tr>
     <td>Deepseek-OCR</td>
     <td>3B</td>
@@ -801,19 +1057,28 @@ To illustrate the impact of real-world visual distortions on document parsing, w
     <td>1.2B</td>
     <td>89.57</td>
     <td>0.065</td>
-    <td><u>88.36</u></td>
+    <td>88.36</td>
     <td>86.87</td>
     <td>0.062</td>
   </tr>
   <tr>
     <td>PaddleOCR-VL </td>
     <td>0.9B</td>
-    <td><u>89.61</u></td>
+    <td>89.61</td>
     <td><u>0.049</u></td>
     <td>86.66</td>
-    <td><u>87.02</u></td>
+    <td>87.02</td>
     <td><u>0.055</u></td>
   </tr>
+  <tr>
+    <td>GLM-OCR</td>
+    <td>0.9B</td>
+    <td><u>91.12</u></td>
+    <td>0.059</td>
+    <td><u>91.02</u></td>
+    <td><u>88.20</u></td>
+    <td>0.071</td>
+  </tr>
   <tr>
     <td><strong>PaddleOCR-VL-1.5</strong></td>
     <td>0.9B</td>
@@ -829,7 +1094,7 @@ To illustrate the impact of real-world visual distortions on document parsing, w
 
 ---
 
-### 5. Skew
+### 6. Skew
 
 <table border="1" style="width:100%; border-collapse: collapse;">
   <tr>
@@ -884,16 +1149,16 @@ To illustrate the impact of real-world visual distortions on document parsing, w
     <td>Qwen2.5-VL-72B</td>
     <td>72B</td>
     <td>86.90</td>
-    <td><u>0.077</u></td>
+    <td>0.077</td>
     <td>87.26</td>
     <td>81.14</td>
-    <td><u>0.091</u></td>
+    <td>0.091</td>
   </tr>
   <tr>
     <td>Gemini-2.5 Pro</td>
     <td>-</td>
     <td>89.07</td>
-    <td><u>0.077</u></td>
+    <td>0.077</td>
     <td>87.89</td>
     <td>86.99</td>
     <td>0.104</td>
@@ -908,7 +1173,7 @@ To illustrate the impact of real-world visual distortions on document parsing, w
     <td>0.092</td>
   </tr>
   <tr>
-    <td rowspan="12">Specialized VLMs</td>
+    <td rowspan="14">Specialized VLMs</td>
     <td>Dolphin-1.5 </td>
     <td>0.3B</td>
     <td>28.16</td>
@@ -926,6 +1191,15 @@ To illustrate the impact of real-world visual distortions on document parsing, w
     <td>33.22</td>
     <td>0.321</td>
   </tr>
+  <tr>
+    <td>Deepseek-OCR 2</td>
+    <td>3B</td>
+    <td>61.28</td>
+    <td>0.295</td>
+    <td>66.16</td>
+    <td>47.18</td>
+    <td>0.221</td>
+  </tr>
   <tr>
     <td>MonkeyOCR-pro-1.2B </td>
     <td>1.9B</td>
@@ -1007,6 +1281,15 @@ To illustrate the impact of real-world visual distortions on document parsing, w
     <td>75.74</td>
     <td>0.094</td>
   </tr>
+  <tr>
+    <td>GLM-OCR</td>
+    <td>0.9B</td>
+    <td>85.39</td>
+    <td>0.099</td>
+    <td>85.78</td>
+    <td>80.28</td>
+    <td>0.156</td>
+  </tr>
   <tr>
     <td><strong>PaddleOCR-VL-1.5</strong></td>
     <td>0.9B</td>
@@ -1031,4 +1314,16 @@ Real5-OmniDocBench is based on the original OmniDocBench dataset and adopts its
 
 ## Citation
 
-If you use Real5-OmniDocBench in your research, please cite our dataset and refer also to the original OmniDocBench paper.
+If you use Real5-OmniDocBench in your research, please cite our dataset and refer also to the original OmniDocBench paper.
+
+```
+@misc{zhou2026real5omnidocbenchfullscalephysicalreconstruction,
+      title={Real5-OmniDocBench: A Full-Scale Physical Reconstruction Benchmark for Robust Document Parsing in the Wild}, 
+      author={Changda Zhou and Ziyue Gao and Xueqing Wang and Tingquan Gao and Cheng Cui and Jing Tang and Yi Liu},
+      year={2026},
+      eprint={2603.04205},
+      archivePrefix={arXiv},
+      primaryClass={cs.CV},
+      url={https://arxiv.org/abs/2603.04205}, 
+}
+```