Anirudh Balaraman commited on
Commit
c0334de
·
unverified ·
1 Parent(s): ed15825

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +20 -8
README.md CHANGED
@@ -35,14 +35,14 @@ Deep learning methods used in medical AI—particularly for csPCa prediction and
35
 
36
 
37
  ## 🚀 Quick Start
38
- 1. Clone and Setup
39
  ```bash
40
  git clone https://github.com/anirudhbalaraman/WSAttention-Prostate.git
41
  cd WSAttention-Prostate
42
  pip install -r requirements.txt
43
  pytest tests/
44
  ```
45
- 2. Model Download
46
 
47
  ```bash
48
  mkdir -p ./models
@@ -52,14 +52,27 @@ curl -L -o models/file3.pth https://huggingface.co/anirudh0410/WSAttention-Prost
52
  ```
53
 
54
  ## 🚀 Usage
 
 
 
 
 
 
 
 
 
 
55
 
56
  ### Preprocessing
57
 
 
58
  ```bash
59
- python preprocess_main.py --config config/config_preprocess.yaml \
60
- --steps register_and_crop get_segmentation_mask histogram_match get_heatmap
 
61
  ```
62
 
 
63
  ### PI-RADS Training
64
 
65
  ```bash
@@ -71,12 +84,11 @@ python run_pirads.py --mode train --config config/config_pirads_train.yaml
71
  ```bash
72
  python run_cspca.py --mode train --config config/config_cspca_train.yaml
73
  ```
74
-
75
- ### Inference
76
 
77
  ```bash
78
- python run_pirads.py --mode test --config config/config_pirads_test.yaml --checkpoint <path>
79
- python run_cspca.py --mode test --config config/config_cspca_test.yaml --checkpoint_cspca <path>
80
  python run_inference.py --config config/config_preprocess.yaml
81
  ```
82
 
 
35
 
36
 
37
  ## 🚀 Quick Start
38
+ ### 1. Clone and Setup
39
  ```bash
40
  git clone https://github.com/anirudhbalaraman/WSAttention-Prostate.git
41
  cd WSAttention-Prostate
42
  pip install -r requirements.txt
43
  pytest tests/
44
  ```
45
+ ### 2. Model Download
46
 
47
  ```bash
48
  mkdir -p ./models
 
52
  ```
53
 
54
  ## 🚀 Usage
55
+ ### Inference
56
+
57
+ Run run_inference.py to execute the full pipeline, from preprocessing to model predictions. The script accepts paths to T2W, DWI, and ADC sequences, as well as an output directory(output_dir), which can be specified in config_preprocess.yaml. ***NOTE:*** For each scan, all sequences should share the same filename, and the input files must be in NRRD format.
58
+
59
+ Outputs are risk of csPCa, PI-RADS score and coordinaates of top 5 salient patches for each scan summarised in results.json saved in output_dir along with the intermediary files from pre processing including the prostate segmentation mask. The patches can be visualised using visualisation.ipynb
60
+
61
+
62
+ ```bash
63
+ python run_inference.py --config config/config_preprocess.yaml
64
+ ```
65
 
66
  ### Preprocessing
67
 
68
+ Execute preprocess_main.py to preprocess your MRI files. Each sequence—T2W, DWI, and ADC—must be placed in separate folders, with paths specified in config_preprocess.yaml.
69
  ```bash
70
+ python preprocess_main.py \
71
+ --steps register_and_crop get_segmentation_mask histogram_match get_heatmap \
72
+ --config config/config_preprocess.yaml
73
  ```
74
 
75
+
76
  ### PI-RADS Training
77
 
78
  ```bash
 
84
  ```bash
85
  python run_cspca.py --mode train --config config/config_cspca_train.yaml
86
  ```
87
+ ### Testing
 
88
 
89
  ```bash
90
+ python run_pirads.py --mode test --config config/config_pirads_test.yaml
91
+ python run_cspca.py --mode test --config config/config_cspca_test.yaml
92
  python run_inference.py --config config/config_preprocess.yaml
93
  ```
94