Update README.md
Browse files
README.md
CHANGED
|
@@ -35,8 +35,6 @@ dataset_info:
|
|
| 35 |
[](https://arxiv.org/abs/2505.16647)
|
| 36 |
📫 For queries, contact: [sushant@simula.no](mailto:sushant@simula.no)
|
| 37 |
|
| 38 |
-
---
|
| 39 |
-
|
| 40 |
## Dataset Summary
|
| 41 |
|
| 42 |
**MedMultiPoints** is a curated, multimodal medical imaging dataset designed for **multi-task learning** in the medical domain—spanning **object detection**, **localization**, and **counting** tasks. It integrates data from **endoscopic** and **microscopic** modalities, reflecting real-world clinical diversity.
|
|
@@ -46,7 +44,6 @@ The dataset is introduced in the paper:
|
|
| 46 |
Presented at **IEEE CBMS 2025, Madrid, Spain.**
|
| 47 |
→ [Project Page & Code](https://github.com/Simula/PointDetectCount)
|
| 48 |
|
| 49 |
-
---
|
| 50 |
|
| 51 |
## Features
|
| 52 |
|
|
@@ -58,8 +55,6 @@ Presented at **IEEE CBMS 2025, Madrid, Spain.**
|
|
| 58 |
- Compatible with **Vision-Language Models (VLMs)** and **instruction-tuned pipelines**
|
| 59 |
- JSON-formatted annotations designed for seamless integration with multimodal training
|
| 60 |
|
| 61 |
-
---
|
| 62 |
-
|
| 63 |
## Data Schema
|
| 64 |
|
| 65 |
Each sample in the dataset contains:
|
|
@@ -77,7 +72,6 @@ Each sample in the dataset contains:
|
|
| 77 |
| `classification` | string | Description of annotation type (e.g., pathological-findings) |
|
| 78 |
| `organ` | string | Target organ: `Lower GI`, `Microscopy`, etc. |
|
| 79 |
|
| 80 |
-
---
|
| 81 |
|
| 82 |
## Supported Tasks
|
| 83 |
|
|
@@ -88,8 +82,6 @@ This dataset supports the following **multi-task** settings:
|
|
| 88 |
- 🔢 **Counting** (object count regression)
|
| 89 |
- 🧠 **Multimodal Instruction-Based Learning**
|
| 90 |
|
| 91 |
-
---
|
| 92 |
-
|
| 93 |
## How to Load
|
| 94 |
|
| 95 |
```python
|
|
@@ -105,7 +97,6 @@ points = sample['points']
|
|
| 105 |
count = sample['count']
|
| 106 |
```
|
| 107 |
|
| 108 |
-
---
|
| 109 |
|
| 110 |
## Example
|
| 111 |
|
|
@@ -123,7 +114,6 @@ count = sample['count']
|
|
| 123 |
}
|
| 124 |
```
|
| 125 |
|
| 126 |
-
---
|
| 127 |
|
| 128 |
## Citation
|
| 129 |
|
|
|
|
| 35 |
[](https://arxiv.org/abs/2505.16647)
|
| 36 |
📫 For queries, contact: [sushant@simula.no](mailto:sushant@simula.no)
|
| 37 |
|
|
|
|
|
|
|
| 38 |
## Dataset Summary
|
| 39 |
|
| 40 |
**MedMultiPoints** is a curated, multimodal medical imaging dataset designed for **multi-task learning** in the medical domain—spanning **object detection**, **localization**, and **counting** tasks. It integrates data from **endoscopic** and **microscopic** modalities, reflecting real-world clinical diversity.
|
|
|
|
| 44 |
Presented at **IEEE CBMS 2025, Madrid, Spain.**
|
| 45 |
→ [Project Page & Code](https://github.com/Simula/PointDetectCount)
|
| 46 |
|
|
|
|
| 47 |
|
| 48 |
## Features
|
| 49 |
|
|
|
|
| 55 |
- Compatible with **Vision-Language Models (VLMs)** and **instruction-tuned pipelines**
|
| 56 |
- JSON-formatted annotations designed for seamless integration with multimodal training
|
| 57 |
|
|
|
|
|
|
|
| 58 |
## Data Schema
|
| 59 |
|
| 60 |
Each sample in the dataset contains:
|
|
|
|
| 72 |
| `classification` | string | Description of annotation type (e.g., pathological-findings) |
|
| 73 |
| `organ` | string | Target organ: `Lower GI`, `Microscopy`, etc. |
|
| 74 |
|
|
|
|
| 75 |
|
| 76 |
## Supported Tasks
|
| 77 |
|
|
|
|
| 82 |
- 🔢 **Counting** (object count regression)
|
| 83 |
- 🧠 **Multimodal Instruction-Based Learning**
|
| 84 |
|
|
|
|
|
|
|
| 85 |
## How to Load
|
| 86 |
|
| 87 |
```python
|
|
|
|
| 97 |
count = sample['count']
|
| 98 |
```
|
| 99 |
|
|
|
|
| 100 |
|
| 101 |
## Example
|
| 102 |
|
|
|
|
| 114 |
}
|
| 115 |
```
|
| 116 |
|
|
|
|
| 117 |
|
| 118 |
## Citation
|
| 119 |
|