Update Readme.md
Browse files
README.md
CHANGED
|
@@ -11,4 +11,42 @@ base_model:
|
|
| 11 |
tags:
|
| 12 |
- biology
|
| 13 |
- medical
|
| 14 |
-
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 11 |
tags:
|
| 12 |
- biology
|
| 13 |
- medical
|
| 14 |
+
---
|
| 15 |
+
|
| 16 |
+
# Welcome to SAMF model [MICCAI' 25]!
|
| 17 |
+
**[MICCAI' 25] From Slices to Volumes: Multi-Scale Fusion of 2D and 3D Features for CT Scan Report Generation**
|
| 18 |
+
|
| 19 |
+
| **Model** | **Bleu1** | **Bleu4** | **RougeL** | **Meteor** | **Bert F1** | **Llama Score** |
|
| 20 |
+
|-----------------------|-----------|-----------|------------|------------|-------------|-----------------|
|
| 21 |
+
| CT2Rep | 0.309 | 0.172 | 0.243 | 0.173 | 0.865 | 6.35 |
|
| 22 |
+
| CT-Chat | 0.395 | - | 0.321 | 0.219 | - | 5.664 |
|
| 23 |
+
| Our Baseline (SAMF) | 0.423 | 0.203 | 0.338 | 0.356 | 0.879 | 6.792 |
|
| 24 |
+
| SAMF + *Ao2D* | **0.440** | **0.261** | **0.417** | **0.417** | **0.889** | **7.165** |
|
| 25 |
+
|
| 26 |
+
|
| 27 |
+
## Introduction
|
| 28 |
+
|
| 29 |
+
|
| 30 |
+
|
| 31 |
+
*Slice Attentive Multimodal Fusion (SAMF)* , a framework that combines the rich, high-resolution information from 2D slices with the spatial coherence of 3D volumetric data. Experimental results demonstrate that our method outperforms existing baseline approaches in both report generation and multiple-choice question answering, highlighting the critical role of multidimensional feature integration.
|
| 32 |
+
|
| 33 |
+
|
| 34 |
+
## Model Description
|
| 35 |
+
|
| 36 |
+
- **Model type:** 3D Medical Report Generation and Visual Question Answering
|
| 37 |
+
- **Language(s) (NLP):** English
|
| 38 |
+
- **License:** apache-2.0
|
| 39 |
+
- **Finetuned from model:** microsoft/Phi-3-mini-4k-instruct
|
| 40 |
+
|
| 41 |
+
### Training Data
|
| 42 |
+
|
| 43 |
+
- **Medical Report Generation and Visual Question Answering:** [ibrahimhamamci/CT-RATE](https://huggingface.co/datasets/ibrahimhamamci/CT-RATE), default subset
|
| 44 |
+
|
| 45 |
+
### Hardware Utilization
|
| 46 |
+
|
| 47 |
+
- **Hardware Type:** 1 × NVIDIA-A100
|
| 48 |
+
- **Hours used** around 16 hours
|
| 49 |
+
|
| 50 |
+
### Evaluation
|
| 51 |
+
|
| 52 |
+
To perform evaluation using this model, please refer to our GitHub repository ([serag-ai/SAMF](https://github.com/serag-ai/SAMF.git)), which provides detailed information on how to use it.
|