upd md

README.md

@@ -34,5 +34,25 @@ cd aal_stats_vol
 ## Mirror
 <https://www.modelscope.cn/datasets/Genius-Society/aal_stats_vol>
 
+The raw data of this dataset locates at `./data` in the above mirror repo, whose details are as follows:
+### Attachments
+1. Original data;
+2. Packages;
+3. Intermediate products;
+4. Source code with the output csv data.
+
+### Steps to Achieve the Goal
+1. Skull Stripping: To extract brain from raw MRI images;
+2. Tissue Segmentation: To segment brain into white matters, grey matters and cerebrospinal fluid (CSF);
+3. Registration: To register the standard space to native space;
+4. Measurement(With mask): With the generated masks, to measure the volumes of 90 ROIs;
+5. Classification: To fill out the shell scripts and run them, to write a Python script to train the model and test with the provided testing dataset.
+
+### Classification details
+0. If everything went well before this, a matrix with a size of 50x90 would be obtained: It means there are 50 samples of which each has 90 features(volumes);
+1. Dataset: 40 Training and 10 Testing;
+2. Training: To feed into the classifier to train a model;
+3. Testing: To test the model and provide the result.
+
 ## Reference
 [1] [Chapter II ‐ Classifying AD patients and normal controls from brain images](https://github.com/Genius-Society/medical_image_computing/blob/ad/README.md)