Add USD detection dataset - 3,746 annotations, 24 classes, 100% Front/Back classified

README.dataset.txt

@@ -0,0 +1,6 @@
+# Front/Back of USD > 2025-12-23 4:05pm
+https://universe.roboflow.com/sg-knxss/front-back-of-usd
+
+Provided by a Roboflow user
+License: CC BY 4.0
+

README.md

@@ -22,18 +22,24 @@ A refined COCO-format dataset for detecting US Dollar currency and classifying w
 ## Dataset Summary
 
 - **Total Images**: 7,360
+- **Total Annotations**: 3,746
 - **Format**: COCO (object detection)
-- **Classes**: 33 (denominations × front/back × authentic/counterfeit)
+- **Classes**: 24 (denominations × front/back × authentic/counterfeit)
+- **Classification Accuracy**: 100% (all Front/Back classified)
 
-| Split | Images |
-|-------|--------|
-| Train | 5,290 |
-| Valid | 1,221 |
-| Test | 849 |
+| Split | Images | Annotations |
+|-------|--------|-------------|
+| Train | 5,290 | 2,738 |
+| Valid | 1,221 | 627 |
+| Test | 849 | 381 |
 
-## Classes
+## Classes (24 total)
 
-### Denomination + Side (18 classes)
+**Breakdown:**
+- **12 Regular USD classes**: Front/Back variants for $1, $5, $10, $20, $50, $100
+- **12 Counterfeit USD classes**: Front/Back variants for $1, $5, $10, $20, $50, $100
+
+### Regular USD (12 classes)
 - `100USD-Front`, `100USD-Back`
 - `50USD-Front`, `50USD-Back`
 - `20USD-Front`, `20USD-Back`
@@ -41,15 +47,38 @@ A refined COCO-format dataset for detecting US Dollar currency and classifying w
 - `5USD-Front`, `5USD-Back`
 - `1USD-Front`, `1USD-Back`
 
-### Counterfeit Detection (15 classes)
-- Counterfeit versions for each denomination
+### Counterfeit USD (12 classes)
+- `Counterfeit 100 USD Front`, `Counterfeit 100 USD Back`
+- `Counterfeit 50USD Front`, `Counterfeit 50USD Back`
+- `Counterfeit 20USD Front`, `Counterfeit 20USD Back`
+- `Counterfeit 10USD Front`, `Counterfeit 10USD Back`
+- `Counterfeit 5USD Front`, `Counterfeit 5USD Back`
+- `Counterfeit 1USD Front`, `Counterfeit 1USD Back`
+
+**Note**: All $2 bills (289) and generic annotations (13) removed - only Front/Back classified data remains.
 
 ## Annotation Refinement
 
-This dataset was refined using Roboflow's `usd-classification/1` model to reclassify generic labels (e.g., `100USD`) into specific front/back variants:
+This dataset was refined using Roboflow's `usd-classification/1` model:
+
+### Phase 1: Regular USD ✅
+- Reclassified 2,236 generic labels to Front/Back variants
+- **97% success rate**
+
+### Phase 2: Counterfeit USD ✅  
+- Reclassified 943 counterfeit annotations across all splits (train/valid/test)
+- **97.8% success rate** (269/275 in valid/test, 674/762 in train)
+- Only 13 annotations remain generic (SSL errors during classification)
 
-- **2,236 annotations** auto-reclassified
-- **97% success rate** on classification
+### Phase 3: Data Cleaning ✅
+- Removed 289 $2 bill annotations (146 regular + 143 counterfeit)
+- Reason: Model lacks "two-front"/"two-back" classes, generalization only 75% accurate
+
+## Final Statistics
+
+- **3,746 annotations** - 100% classified to Front/Back
+- **24 classes** - 12 regular + 12 counterfeit
+- **0 $2 bills** (all 289 removed - 146 regular + 143 counterfeit)
 
 ## Usage
 
@@ -59,12 +88,15 @@ from datasets import load_dataset
 dataset = load_dataset("ebowwa/usd-side-coco-annotations")
 ```
 
-Or download the zip directly and extract for use with YOLO/RF-DETR training.
+Or download directly and extract for use with YOLO/RF-DETR training.
 
 ## Source
 
-Original dataset from [Roboflow](https://app.roboflow.com) - "Front/Back of USD" project.
+Original dataset from [Roboflow](https://app.roboflow.com/ds/OWTzlaSd01?key=LTqYj5YlFh) - "Front/Back of USD 2" project.
+
+Refined using automated Roboflow classification API with incremental saving for fault tolerance.
 
 ## License
 
 MIT
+

README.roboflow.txt

@@ -0,0 +1,27 @@
+
+Front/Back of USD - v2 2025-12-23 4:05pm
+==============================
+
+This dataset was exported via roboflow.com on December 24, 2025 at 12:20 AM GMT
+
+Roboflow is an end-to-end computer vision platform that helps you
+* collaborate with your team on computer vision projects
+* collect & organize images
+* understand and search unstructured image data
+* annotate, and create datasets
+* export, train, and deploy computer vision models
+* use active learning to improve your dataset over time
+
+For state of the art Computer Vision training notebooks you can use with this dataset,
+visit https://github.com/roboflow/notebooks
+
+To find over 100k other datasets and pre-trained models, visit https://universe.roboflow.com
+
+The dataset includes 7360 images.
+USD-Dollar-cash-counting--c-TiEZ are annotated in COCO format.
+
+The following pre-processing was applied to each image:
+
+No image augmentation techniques were applied.
+
+