Update main.py description and usage instructions; remove unused image asset

main.py

@@ -31,14 +31,27 @@ else:
 gr.set_static_paths(["assets"])
 
 
-description = """Upload a biomedical image and enter prompts (separated by commas) to detect specific features.
+description = """\
+This Space is based on the [BiomedParse model](https://microsoft.github.io/BiomedParse/).
+
+BiomedParse is a model that can detect various targets like organs, diseases, and more
+in biomedical images. The biomedical images can be of different types like CT, MRI, X-ray, etc.
+
+> Note: Don't use this model for medical diagnosis. Always consult a healthcare professional for medical advice.
 
-The model understands these prompts:
-![gpt4_ontology_hierarchy.png](file/assets/gpt4_ontology_hierarchy.png)
+## How to use this demo
 
-Above figure is from the [BiomedParse paper](https://arxiv.org/abs/2405.12971).
+1. Upload a biomedical image
+2. Select the modality type
+3. Select the targets you want to detect
+4. Click on the 'Submit' button to see the prediction
 
-The model understands these types of biomedical images:
+The model will highlight the detected targets in the image and show the targets that were not found below the image.
+Each target is represented by a different color.
+Each target comes with a p-value. A target whose p-value is below 0.05 is considered "not found".
+For more details, check out the paper [BiomedParse: a biomedical foundation model for image parsing of everything everywhere all at once](https://arxiv.org/abs/2405.12971).
+
+## Modality Types
 
 - [Computed Tomography (CT)](https://en.wikipedia.org/wiki/Computed_tomography)
 - [Magnetic Resonance Imaging (MRI)](https://en.wikipedia.org/wiki/Magnetic_resonance_imaging)
@@ -50,7 +63,6 @@ The model understands these types of biomedical images:
 - [Endoscopy](https://en.wikipedia.org/wiki/Endoscopy)
 - [Optical Coherence Tomography (OCT)](https://en.wikipedia.org/wiki/Optical_coherence_tomography)
 
-This Space is based on the [BiomedParse model](https://microsoft.github.io/BiomedParse/).
 """