Datasets:
Commit ·
a0df82d
0
Parent(s):
Duplicate from abdoelsayed/CORU
Browse filesCo-authored-by: Abdelrahman Abdallah <abdoelsayed@users.noreply.huggingface.co>
- .gitattributes +57 -0
- IE/test.csv +0 -0
- IE/train.csv +0 -0
- IE/val.csv +0 -0
- OCR/test.zip +3 -0
- OCR/train.zip +3 -0
- OCR/val.zip +3 -0
- QA/test.zip +3 -0
- README.md +271 -0
- Receipt/labels.txt +3 -0
- Receipt/test.json +3 -0
- Receipt/test.zip +3 -0
- Receipt/train.json +3 -0
- Receipt/train.zip +3 -0
- Receipt/val.json +3 -0
- Receipt/val.zip +3 -0
- images/1.jpg +3 -0
- images/2.jpg +3 -0
- images/3.jpg +3 -0
.gitattributes
ADDED
|
@@ -0,0 +1,57 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
*.7z filter=lfs diff=lfs merge=lfs -text
|
| 2 |
+
*.arrow filter=lfs diff=lfs merge=lfs -text
|
| 3 |
+
*.bin filter=lfs diff=lfs merge=lfs -text
|
| 4 |
+
*.bz2 filter=lfs diff=lfs merge=lfs -text
|
| 5 |
+
*.ckpt filter=lfs diff=lfs merge=lfs -text
|
| 6 |
+
*.ftz filter=lfs diff=lfs merge=lfs -text
|
| 7 |
+
*.gz filter=lfs diff=lfs merge=lfs -text
|
| 8 |
+
*.h5 filter=lfs diff=lfs merge=lfs -text
|
| 9 |
+
*.joblib filter=lfs diff=lfs merge=lfs -text
|
| 10 |
+
*.lfs.* filter=lfs diff=lfs merge=lfs -text
|
| 11 |
+
*.lz4 filter=lfs diff=lfs merge=lfs -text
|
| 12 |
+
*.mlmodel filter=lfs diff=lfs merge=lfs -text
|
| 13 |
+
*.model filter=lfs diff=lfs merge=lfs -text
|
| 14 |
+
*.msgpack filter=lfs diff=lfs merge=lfs -text
|
| 15 |
+
*.npy filter=lfs diff=lfs merge=lfs -text
|
| 16 |
+
*.npz filter=lfs diff=lfs merge=lfs -text
|
| 17 |
+
*.onnx filter=lfs diff=lfs merge=lfs -text
|
| 18 |
+
*.ot filter=lfs diff=lfs merge=lfs -text
|
| 19 |
+
*.parquet filter=lfs diff=lfs merge=lfs -text
|
| 20 |
+
*.pb filter=lfs diff=lfs merge=lfs -text
|
| 21 |
+
*.pickle filter=lfs diff=lfs merge=lfs -text
|
| 22 |
+
*.pkl filter=lfs diff=lfs merge=lfs -text
|
| 23 |
+
*.pt filter=lfs diff=lfs merge=lfs -text
|
| 24 |
+
*.pth filter=lfs diff=lfs merge=lfs -text
|
| 25 |
+
*.rar filter=lfs diff=lfs merge=lfs -text
|
| 26 |
+
*.safetensors filter=lfs diff=lfs merge=lfs -text
|
| 27 |
+
saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
| 28 |
+
*.tar.* filter=lfs diff=lfs merge=lfs -text
|
| 29 |
+
*.tar filter=lfs diff=lfs merge=lfs -text
|
| 30 |
+
*.tflite filter=lfs diff=lfs merge=lfs -text
|
| 31 |
+
*.tgz filter=lfs diff=lfs merge=lfs -text
|
| 32 |
+
*.wasm filter=lfs diff=lfs merge=lfs -text
|
| 33 |
+
*.xz filter=lfs diff=lfs merge=lfs -text
|
| 34 |
+
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 35 |
+
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 36 |
+
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
| 37 |
+
# Audio files - uncompressed
|
| 38 |
+
*.pcm filter=lfs diff=lfs merge=lfs -text
|
| 39 |
+
*.sam filter=lfs diff=lfs merge=lfs -text
|
| 40 |
+
*.raw filter=lfs diff=lfs merge=lfs -text
|
| 41 |
+
# Audio files - compressed
|
| 42 |
+
*.aac filter=lfs diff=lfs merge=lfs -text
|
| 43 |
+
*.flac filter=lfs diff=lfs merge=lfs -text
|
| 44 |
+
*.mp3 filter=lfs diff=lfs merge=lfs -text
|
| 45 |
+
*.ogg filter=lfs diff=lfs merge=lfs -text
|
| 46 |
+
*.wav filter=lfs diff=lfs merge=lfs -text
|
| 47 |
+
# Image files - uncompressed
|
| 48 |
+
*.bmp filter=lfs diff=lfs merge=lfs -text
|
| 49 |
+
*.gif filter=lfs diff=lfs merge=lfs -text
|
| 50 |
+
*.png filter=lfs diff=lfs merge=lfs -text
|
| 51 |
+
*.tiff filter=lfs diff=lfs merge=lfs -text
|
| 52 |
+
# Image files - compressed
|
| 53 |
+
*.jpg filter=lfs diff=lfs merge=lfs -text
|
| 54 |
+
*.jpeg filter=lfs diff=lfs merge=lfs -text
|
| 55 |
+
*.webp filter=lfs diff=lfs merge=lfs -text
|
| 56 |
+
*.json filter=lfs diff=lfs merge=lfs -text
|
| 57 |
+
*.txt filter=lfs diff=lfs merge=lfs -text
|
IE/test.csv
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
IE/train.csv
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
IE/val.csv
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
OCR/test.zip
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:724f93707d135e861c63dbb4743ff11b187820273f3de6b7f7a9050ed4816f2c
|
| 3 |
+
size 38735132
|
OCR/train.zip
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5d19042b03f1e298a073d21f55749caffd692b6747ecff0aa41fc5974b95cb74
|
| 3 |
+
size 181347200
|
OCR/val.zip
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5a8f9780ef40053c9cf7d0bd43643b5784882d8d7bba34fefebab85593b60997
|
| 3 |
+
size 39074274
|
QA/test.zip
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f3a71f126369c7a18381846105d759c51f7058b2ccb6862c051a6e1f902e79b2
|
| 3 |
+
size 3927168240
|
README.md
ADDED
|
@@ -0,0 +1,271 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: mit
|
| 3 |
+
task_categories:
|
| 4 |
+
- object-detection
|
| 5 |
+
- text-classification
|
| 6 |
+
- zero-shot-classification
|
| 7 |
+
language:
|
| 8 |
+
- en
|
| 9 |
+
- ar
|
| 10 |
+
size_categories:
|
| 11 |
+
- 10K<n<100K
|
| 12 |
+
---
|
| 13 |
+
# ReceiptSense: Beyond Traditional OCR - A Dataset for Receipt Understanding
|
| 14 |
+
|
| 15 |
+
[](https://arxiv.org/abs/2406.04493)
|
| 16 |
+
[](https://huggingface.co/datasets/abdoelsayed/CORU)
|
| 17 |
+
[]()
|
| 18 |
+
|
| 19 |
+
## 🔥 News
|
| 20 |
+
- **[2024]** ReceiptSense dataset is now publicly available!
|
| 21 |
+
- **[2024]** Paper accepted and published
|
| 22 |
+
|
| 23 |
+
## 📖 Abstract
|
| 24 |
+
|
| 25 |
+
Multilingual OCR and information extraction from receipts remains challenging, particularly for complex scripts like Arabic. We introduce **ReceiptSense**, a comprehensive dataset designed for Arabic-English receipt understanding comprising:
|
| 26 |
+
|
| 27 |
+
- **20,000** annotated receipts from diverse retail settings
|
| 28 |
+
- **30,000** OCR-annotated images
|
| 29 |
+
- **10,000** item-level annotations
|
| 30 |
+
- **1,265** receipt images with **40 question-answer pairs each** for Receipt QA
|
| 31 |
+
|
| 32 |
+
The dataset captures merchant names, item descriptions, prices, receipt numbers, and dates to support object detection, OCR, information extraction, and question-answering tasks. We establish baseline performance using traditional methods (Tesseract OCR) and advanced neural networks, demonstrating the dataset's effectiveness for processing complex, noisy real-world receipt layouts.
|
| 33 |
+
|
| 34 |
+
## 🎯 Key Features
|
| 35 |
+
|
| 36 |
+
### ✨ **Multilingual Support**
|
| 37 |
+
- **Arabic-English** bilingual receipts
|
| 38 |
+
- Real-world mixed-language content
|
| 39 |
+
- Complex script handling for Arabic text
|
| 40 |
+
|
| 41 |
+
### 📊 **Comprehensive Annotations**
|
| 42 |
+
- **Object Detection**: Bounding boxes for key receipt elements
|
| 43 |
+
- **OCR**: Character and word-level text recognition
|
| 44 |
+
- **Information Extraction**: Structured data extraction
|
| 45 |
+
- **Receipt QA**: Question-answering capabilities
|
| 46 |
+
|
| 47 |
+
### 🏪 **Diverse Retail Environments**
|
| 48 |
+
- Supermarkets and grocery stores
|
| 49 |
+
- Restaurants and cafes
|
| 50 |
+
- Clothing and retail shops
|
| 51 |
+
- Various geographical regions
|
| 52 |
+
|
| 53 |
+
### 🔧 **Real-world Challenges**
|
| 54 |
+
- Noisy and degraded image quality
|
| 55 |
+
- Complex receipt layouts
|
| 56 |
+
- Mixed fonts and orientations
|
| 57 |
+
- Authentic retail scenarios
|
| 58 |
+
|
| 59 |
+
## 📈 Dataset Statistics
|
| 60 |
+
|
| 61 |
+
| Component | Training | Validation | Test | Total |
|
| 62 |
+
|-----------|----------|------------|------|-------|
|
| 63 |
+
| **Key Information Detection** | 12,600 | 3,700 | 3,700 | **20,000** |
|
| 64 |
+
| **OCR Dataset** | 21,000 | 4,500 | 4,500 | **30,000** |
|
| 65 |
+
| **Item Information Extraction** | 7,000 | 1,500 | 1,500 | **10,000** |
|
| 66 |
+
| **Receipt QA** | - | - | 1,265 | **1,265** |
|
| 67 |
+
|
| 68 |
+
### Language Distribution
|
| 69 |
+
- **Arabic**: 53.6%
|
| 70 |
+
- **English**: 26.2%
|
| 71 |
+
- **Mixed Language**: 20.3%
|
| 72 |
+
|
| 73 |
+
### Receipt QA Coverage
|
| 74 |
+
- **Merchant/Payment/Date Metadata**: 30%
|
| 75 |
+
- **Item-level Information**: 50%
|
| 76 |
+
- **Tax/Total/Payment Details**: 20%
|
| 77 |
+
|
| 78 |
+
## 🖼️ Sample Images
|
| 79 |
+
|
| 80 |
+
<div align="center">
|
| 81 |
+
|
| 82 |
+
| Sample 1 | Sample 2 | Sample 3 | Sample 4 | Sample 5 |
|
| 83 |
+
|----------|----------|----------|----------|----------|
|
| 84 |
+
| <img src="images/0cf392e3-e6bf-4bd7-85d5-7f91c73cdcaf.jpg" width="150" height="200"> | <img src="images/0dccefa6-6928-499e-8aae-15c04d18cc94.jpg" width="150" height="200"> | <img src="images/0dd4ada2-681e-42e7-b398-e093bc8b81c3.jpg" width="150" height="200"> | <img src="images/0ef51dc7-4a0a-47e6-bc59-41f609d1c98d.jpg" width="150" height="200"> | <img src="images/0f369dc1-1c5b-41b1-97bc-c9b94d53cd40.jpg" width="150" height="200"> |
|
| 85 |
+
|
| 86 |
+
*Examples of annotated receipt images showcasing the variety of formats, languages, and complex text layouts*
|
| 87 |
+
|
| 88 |
+
</div>
|
| 89 |
+
|
| 90 |
+
## 🎯 Supported Tasks
|
| 91 |
+
|
| 92 |
+
### 1. 🎯 **Key Information Detection**
|
| 93 |
+
Extract essential receipt information including:
|
| 94 |
+
- Merchant names
|
| 95 |
+
- Transaction dates
|
| 96 |
+
- Receipt numbers
|
| 97 |
+
- Item lists and descriptions
|
| 98 |
+
- Total amounts
|
| 99 |
+
|
| 100 |
+
### 2. 🔍 **OCR (Optical Character Recognition)**
|
| 101 |
+
Box-level text annotations for:
|
| 102 |
+
- Multilingual text recognition
|
| 103 |
+
- Complex layout understanding
|
| 104 |
+
- Noisy image processing
|
| 105 |
+
|
| 106 |
+
### 3. 📝 **Information Extraction**
|
| 107 |
+
Detailed item-level analysis:
|
| 108 |
+
- Item names and descriptions
|
| 109 |
+
- Prices and quantities
|
| 110 |
+
- Categories and classifications
|
| 111 |
+
- Brands and packaging information
|
| 112 |
+
|
| 113 |
+
### 4. ❓ **Receipt Question Answering**
|
| 114 |
+
Comprehensive QA capabilities covering:
|
| 115 |
+
- Receipt metadata queries
|
| 116 |
+
- Item-specific questions
|
| 117 |
+
- Transaction summary questions
|
| 118 |
+
- Payment and tax information
|
| 119 |
+
|
| 120 |
+
## 📥 Download Links
|
| 121 |
+
|
| 122 |
+
### 🎯 Key Information Detection
|
| 123 |
+
- **Training Set**: [Download (12.6K images)](https://huggingface.co/datasets/abdoelsayed/CORU/resolve/main/Receipt/train.zip?download=true)
|
| 124 |
+
- **Validation Set**: [Download (3.7K images)](https://huggingface.co/datasets/abdoelsayed/CORU/resolve/main/Receipt/val.zip?download=true)
|
| 125 |
+
- **Test Set**: [Download (3.7K images)](https://huggingface.co/datasets/abdoelsayed/CORU/resolve/main/Receipt/test.zip?download=true)
|
| 126 |
+
|
| 127 |
+
### 🔍 OCR Dataset
|
| 128 |
+
- **Training Set**: [Download (21K images)](https://huggingface.co/datasets/abdoelsayed/CORU/resolve/main/OCR/train.zip?download=true)
|
| 129 |
+
- **Validation Set**: [Download (4.5K images)](https://huggingface.co/datasets/abdoelsayed/CORU/resolve/main/OCR/val.zip?download=true)
|
| 130 |
+
- **Test Set**: [Download (4.5K images)](https://huggingface.co/datasets/abdoelsayed/CORU/resolve/main/OCR/test.zip?download=true)
|
| 131 |
+
|
| 132 |
+
### 📝 Item Information Extraction
|
| 133 |
+
- **Training Set**: [Download (7K items)](https://huggingface.co/datasets/abdoelsayed/CORU/resolve/main/IE/train.csv?download=true)
|
| 134 |
+
- **Validation Set**: [Download (1.5K items)](https://huggingface.co/datasets/abdoelsayed/CORU/resolve/main/IE/val.csv?download=true)
|
| 135 |
+
- **Test Set**: [Download (1.5K items)](https://huggingface.co/datasets/abdoelsayed/CORU/resolve/main/IE/test.csv?download=true)
|
| 136 |
+
|
| 137 |
+
### ❓ Receipt Question Answering
|
| 138 |
+
- **Test Set**: [Download (1,265 receipts with 50.6K QA pairs)](https://huggingface.co/datasets/abdoelsayed/CORU/resolve/main/QA/test.zip?download=true)
|
| 139 |
+
|
| 140 |
+
> ⚠️ **Note**: All receipt datasets have been updated to include PII-redacted versions for privacy protection.
|
| 141 |
+
|
| 142 |
+
## 🏆 Baseline Results
|
| 143 |
+
|
| 144 |
+
### Object Detection Performance
|
| 145 |
+
| Model | Backbone | Precision | Recall | mAP50 | mAP50-95 |
|
| 146 |
+
|-------|----------|-----------|--------|-------|----------|
|
| 147 |
+
| **YOLOv7** | - | **76.0%** | **85.6%** | **79.2%** | 43.7% |
|
| 148 |
+
| YOLOv8 | - | 74.6% | 81.0% | 76.1% | 45.3% |
|
| 149 |
+
| YOLOv9 | - | 75.7% | 83.4% | 77.9% | **46.7%** |
|
| 150 |
+
| DINO | Swin-T | - | - | - | **32.2%** (Avg IoU) |
|
| 151 |
+
|
| 152 |
+
### OCR Performance
|
| 153 |
+
| Model | CER ↓ | WER ↓ |
|
| 154 |
+
|-------|-------|-------|
|
| 155 |
+
| Tesseract | 15.56% | 30.78% |
|
| 156 |
+
| Attention-Gated CNN-BiGRU | 14.85% | 27.22% |
|
| 157 |
+
| Our OCR Model | 7.83% | 27.24% |
|
| 158 |
+
| **Azura OCR** | **6.39%** | **25.97%** |
|
| 159 |
+
|
| 160 |
+
### Receipt QA Performance
|
| 161 |
+
| Model | Precision | Recall | Exact Match | Contains |
|
| 162 |
+
|-------|-----------|--------|-------------|----------|
|
| 163 |
+
| **GPT-4o** | **37.7%** | **36.4%** | **35.0%** | **29.1%** |
|
| 164 |
+
| Llama3.2 (11B) | 32.6% | 31.3% | 31.6% | 25.9% |
|
| 165 |
+
| Phi3.5 | 28.4% | 29.1% | 28.8% | 23.7% |
|
| 166 |
+
| Internvl2 (8B) | 24.2% | 23.8% | 23.1% | 19.4% |
|
| 167 |
+
|
| 168 |
+
## 🚀 Getting Started
|
| 169 |
+
|
| 170 |
+
### Quick Start
|
| 171 |
+
```python
|
| 172 |
+
# Install required packages
|
| 173 |
+
pip install datasets transformers torch
|
| 174 |
+
|
| 175 |
+
# Load the dataset
|
| 176 |
+
from datasets import load_dataset
|
| 177 |
+
|
| 178 |
+
# Load Receipt QA dataset
|
| 179 |
+
qa_dataset = load_dataset("abdoelsayed/CORU", "qa")
|
| 180 |
+
|
| 181 |
+
# Load OCR dataset
|
| 182 |
+
ocr_dataset = load_dataset("abdoelsayed/CORU", "ocr")
|
| 183 |
+
|
| 184 |
+
# Load Information Extraction dataset
|
| 185 |
+
ie_dataset = load_dataset("abdoelsayed/CORU", "ie")
|
| 186 |
+
```
|
| 187 |
+
|
| 188 |
+
### Dataset Structure
|
| 189 |
+
```
|
| 190 |
+
ReceiptSense/
|
| 191 |
+
├── Receipt/ # Key Information Detection
|
| 192 |
+
│ ├── images/ # Receipt images
|
| 193 |
+
│ └── annotations/ # YOLO/COCO format annotations
|
| 194 |
+
├── OCR/ # OCR Dataset
|
| 195 |
+
│ ├── images/ # Text line images
|
| 196 |
+
│ └── labels/ # Character annotations
|
| 197 |
+
├── IE/ # Information Extraction
|
| 198 |
+
│ └── data.csv # Structured item data
|
| 199 |
+
└── QA/ # Receipt Question Anshwering
|
| 200 |
+
├── images/ # Receipt images
|
| 201 |
+
└── qa_pairs.json # Question-answer pairs
|
| 202 |
+
```
|
| 203 |
+
|
| 204 |
+
## 🔬 Applications
|
| 205 |
+
|
| 206 |
+
- **💳 Expense Management**: Automated expense tracking and categorization
|
| 207 |
+
- **📦 Inventory Management**: Real-time inventory updates from receipt data
|
| 208 |
+
- **🏪 Retail Analytics**: Customer behavior and purchasing pattern analysis
|
| 209 |
+
- **🤖 Document AI**: Multilingual document understanding systems
|
| 210 |
+
- **📱 Mobile Apps**: Receipt scanning and digitization applications
|
| 211 |
+
|
| 212 |
+
## 🤝 Comparison with Existing Datasets
|
| 213 |
+
|
| 214 |
+
| Dataset | Images | Categories | Languages | Item IE | Receipt QA | Year |
|
| 215 |
+
|---------|--------|------------|-----------|---------|------------|------|
|
| 216 |
+
| SROIE | 1,000 | 4 | English | ✓ | ✗ | 2019 |
|
| 217 |
+
| CORD | 1,000 | 8 | English | ✓ | ✗ | 2019 |
|
| 218 |
+
| MC-OCR | 2,436 | 4 | EN + Vietnamese | ✓ | ✗ | 2021 |
|
| 219 |
+
| UIT | 2,147 | 4 | EN + Vietnamese | ✓ | ✗ | 2022 |
|
| 220 |
+
| **ReceiptSense** | **20,000** | **5** | **Arabic + English** | **✓** | **✓** | **2024** |
|
| 221 |
+
|
| 222 |
+
## 🏛️ Ethics and Privacy
|
| 223 |
+
|
| 224 |
+
- All receipts collected with explicit user consent through the DISCO application
|
| 225 |
+
- Comprehensive 4-step PII redaction process implemented
|
| 226 |
+
- Privacy protocols strictly followed during data collection
|
| 227 |
+
- Independent verification and cross-checking procedures
|
| 228 |
+
|
| 229 |
+
## 👥 Authors
|
| 230 |
+
|
| 231 |
+
**Abdelrahman Abdallah¹**, **Mahmoud Abdalla²**, **Mahmoud SalahEldin Kasem²**, **Mohamed Mahmoud²**, **Ibrahim Abdelhalim³**, **Mohamed Elkasaby⁴**, **Yasser Elbendary⁴**, **Adam Jatowt¹**
|
| 232 |
+
|
| 233 |
+
¹University of Innsbruck, Innsbruck, Tyrol, Austria
|
| 234 |
+
²Chungbuk National University, Cheongju, Republic of Korea
|
| 235 |
+
³University of Louisville, Louisville, USA
|
| 236 |
+
⁴DISCO, Cairo, Egypt
|
| 237 |
+
|
| 238 |
+
## 📚 Citation
|
| 239 |
+
|
| 240 |
+
If you find ReceiptSense useful for your research, please consider citing our paper:
|
| 241 |
+
|
| 242 |
+
```bibtex
|
| 243 |
+
@article{abdallah2024receiptsense,
|
| 244 |
+
title={ReceiptSense: Beyond Traditional OCR - A Dataset for Receipt Understanding},
|
| 245 |
+
author={Abdelrahman Abdallah and Mahmoud Abdalla and Mahmoud SalahEldin Kasem and Mohamed Mahmoud and Ibrahim Abdelhalim and Mohamed Elkasaby and Yasser Elbendary and Adam Jatowt},
|
| 246 |
+
year={2024},
|
| 247 |
+
journal={ACM Conference Proceedings},
|
| 248 |
+
note={Comprehensive multilingual receipt understanding dataset}
|
| 249 |
+
}
|
| 250 |
+
```
|
| 251 |
+
|
| 252 |
+
## 📄 License
|
| 253 |
+
|
| 254 |
+
This dataset is released under the MIT License. See [LICENSE](LICENSE) file for details.
|
| 255 |
+
|
| 256 |
+
## 🔗 Links
|
| 257 |
+
|
| 258 |
+
- 📄 **Paper**: [arXiv:2406.04493](https://arxiv.org/abs/2406.04493)
|
| 259 |
+
- 🤗 **HuggingFace**: [abdoelsayed/CORU](https://huggingface.co/datasets/abdoelsayed/CORU)
|
| 260 |
+
- 💼 **DISCO App**: [https://discoapp.ai/](https://discoapp.ai/)
|
| 261 |
+
- 📧 **Contact**: [abdelrahman.abdallah@uibk.ac.at](mailto:abdelrahman.abdallah@uibk.ac.at)
|
| 262 |
+
|
| 263 |
+
---
|
| 264 |
+
|
| 265 |
+
<div align="center">
|
| 266 |
+
|
| 267 |
+
**🌟 Star this repository if you find it helpful! 🌟**
|
| 268 |
+
|
| 269 |
+

|
| 270 |
+
|
| 271 |
+
</div>
|
Receipt/labels.txt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4d7ee3c5720620f15791980c30514e110a89c64f40c6b1f9a296b2523ba46555
|
| 3 |
+
size 31
|
Receipt/test.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f9bd21061515ca79ce1ceecf0837faa8c1f418eaa406fc6d38c6eff012ee6ab7
|
| 3 |
+
size 5022111
|
Receipt/test.zip
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:351d1c5a2ab1ef12399787679563fbc9922a8d18e0f6296e16d7db0e0a4f7f41
|
| 3 |
+
size 1057581808
|
Receipt/train.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:40b3af75daa88bc9f6cefe4681d4894a1a544dccdeddba98e2a1bccb6d2537b7
|
| 3 |
+
size 10816731
|
Receipt/train.zip
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5b90cce9f5f0075b0546b95256a2de4bebf539a9a30a56f64263db9ef7521f25
|
| 3 |
+
size 9391051323
|
Receipt/val.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:17989a121c88eecb97d1f260c978aab747b1dba3487897b00a5d2baa829ab1b6
|
| 3 |
+
size 1090261
|
Receipt/val.zip
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2d9c0736a80a4ca7ce874672b2198d8780eff28595c6a832005229c10792cbff
|
| 3 |
+
size 1245238944
|
images/1.jpg
ADDED
|
Git LFS Details
|
images/2.jpg
ADDED
|
Git LFS Details
|
images/3.jpg
ADDED
|
Git LFS Details
|