SbNet Signboard Detection and Classification Dataset

<h2>Phase 1: Signboard Detection Dataset</h2>
<i>This phase focuses on detecting signboards in street images.</i><br>

- **Total Images:** 8,366<br>
- **Image Format:** JPG (8,366 images)<br>
- **Resolution:**<br>
- Minimum: (720, 443)<br>
- Maximum: (9,280, 8,285)<br>
- Mean: (4,202, 3,138)<br>
- Median: (4,032, 3,024)<br>
- **Aspect Ratio:**<br>
- Minimum: 0.5625<br>
- Maximum: 5.7043<br>
- Mean: 1.3691<br>
- Most Frequent: 1.3333<br>
- Standard Deviation: 0.2329<br>
- **File Size (KB):**<br>
- Minimum: 88.19 KB<br>
- Maximum: 41,266.50 KB<br>
- Mean: 5,796.19 KB<br>
- Total Dataset Size: 48,490,924.91 KB<br>
- **Color Statistics:**<br>
- Color Mode: RGB (8,366 images)<br>
- Mean Color (RGB): (110.32, 112.77, 118.16)<br>
- Standard Deviation (RGB): (65.71, 65.36, 65.82)<br>
- **Brightness:**<br>
- Average: 114.10<br>

---<br>

<h2>Phase 2: Region of Text Interest (RTI) Detection Dataset</h2>
<i>This phase focuses on detecting specific text regions (names and addresses) within signboards.</i><br>

- **Total Images:** 8,036<br>
- **Image Format:** JPG (8,036 images)<br>
- **Resolution:**<br>
- Minimum: (552, 156)<br>
- Maximum: (9,228, 4,682)<br>
- Mean: (2,753, 808)<br>
- Median: (2,741, 781)<br>
- **Aspect Ratio:**<br>
- Minimum: 0.9615<br>
- Maximum: 11.3835<br>
- Mean: 3.6058<br>
- Most Frequent: 4.0<br>
- Standard Deviation: 1.2475<br>
- **File Size (KB):**<br>
- Minimum: 40.54 KB<br>
- Maximum: 7,968.94 KB<br>
- Mean: 653.67 KB<br>
- Total Dataset Size: 5,252,868.26 KB<br>
- **Color Statistics:**<br>
- Color Mode: RGB (8,036 images)<br>
- Mean Color (RGB): (137.58, 136.29, 144.00)<br>
- Standard Deviation (RGB): (47.26, 49.73, 50.89)<br>
- **Brightness:**<br>
- Average: 138.74<br>

---<br>
<br>
<h2>Named Entity Recognition (NER) Dataset</h2>
<i>This dataset is used for categorizing extracted text from signboards.</i><br>
<br>
- **Total Entries:** 42,547<br>
- **Unique Categories:** 10<br>
- **Category Distribution:**<br>
- Religious Sites: 10,641<br>
- Retail Outlets: 8,275<br>
- Educational Institutions: 6,826<br>
- Healthcare Institutions: 4,708<br>
- Restaurants: 3,868<br>
- Pharmacies: 3,637<br>
- Parks: 1,547<br>
- Banks: 1,121<br>
- Stations: 1,094<br>
- Hotels: 830<br>
<br>
#### **Word Count Statistics:**<br>
- **Overall Word Count:**<br>
- Maximum: 18<br>
- Minimum: 1<br>
- Mean: 3.82<br>
- **Category-Wise Word Count:*<br>*
- **Banks:** Mean: 4.65, Max: 11, Min: 1<br>
- **Educational Institutions:** Mean: 4.60, Max: 18, Min: 1<br>
- **Healthcare Institutions:** Mean: 4.02, Max: 16, Min: 1<br>
- **Religious Sites:** Mean: 4.36, Max: 17, Min: 1<br>
- **Retail Outlets:** Mean: 3.08, Max: 15, Min: 1<br>
- **Restaurants:** Mean: 3.36, Max: 13, Min: 1<br>
- **Pharmacies:** Mean: 2.91, Max: 13, Min: 1<br>
- **Parks:** Mean: 3.10, Max: 11, Min: 1<br>
- **Stations:** Mean: 3.72, Max: 17, Min: 1<br>
- **Hotels:** Mean: 3.12, Max: 12, Min: 1<br>

This dataset is structured for a two-phase object detection pipeline with an additional text classification task to categorize extracted text from detected regions.