diff --git a/src/Dataset/dataset/jpg2nii_data_convert.py b/src/Dataset/dataset/jpg2nii_data_convert.py
new file mode 100644
index 0000000..924b4e1
--- /dev/null
+++ b/src/Dataset/dataset/jpg2nii_data_convert.py
@@ -0,0 +1,113 @@
+#processed cases accoring to case_id_list, and save a csv file, with image path and image caption
+import os
+import cv2
+import csv
+import json
+import subprocess
+import pandas as pd
+import numpy as np
+import SimpleITK as sitk
+from tqdm import tqdm
+from collections import defaultdict
+
+def get_image(single_image_dir,single_image_filenames):
+    # single_image_filenames
+    single_image_filenames.sort(key=lambda x: int(x.split('.')[0]))
+    image_list = []
+    for image_filename in single_image_filenames:
+        image_file = os.path.join(single_image_dir, image_filename)
+        #read jpeg to 2D array
+        image_array = cv2.imread(image_file,0)
+        if image_array is not None:
+            image_size = image_array.shape
+            image_array = cv2.resize(image_array,(512,512),interpolation = cv2.INTER_LINEAR)
+            image_list.append(image_array)
+        else:
+            pass
+    image_array = np.array(image_list) #c,w,h
+    if len(image_array.shape) == 3:
+        if image_array.shape[0] < image_array.shape[1]:
+            image_array = image_array.transpose((1, 2, 0))
+        # image_array = np.transpose(image_array, (2,0,1)) # w,h,c
+    return image_array
+
+gray_list = ['CT','MRI','X-ray','Ultrasound','Mammography']
+
+def convert_case(case_id,image_root_dir,json_root_dir,save_case_dict,save_root_dir=None):
+    # save_image_dir 
+    case_images_dir = os.path.join(image_root_dir, case_id)
+    case_json_path = os.path.join(json_root_dir, case_id+'.json')
+    with open(case_json_path, 'r') as f:
+        data = json.load(f)
+    image_nums = (len(data.keys())-1)//2
+    for image_num in range(1,image_nums+1):
+        case_dict = defaultdict(list)
+        image_dir = os.path.join(case_images_dir, str(image_num)) #./images/1/1
+        image_caption = data[str(image_num) + '详情']
+        image_modality = data[str(image_num)][0]['modality']
+        
+        single_image_names = os.listdir(image_dir)
+        single_image_names.sort(key=lambda x: int(x.split('_')[1]))
+        save_image_series = []
+        
+        for single_image_name in single_image_names:
+            single_image_dir = os.path.join(image_dir, single_image_name)
+            
+            save_npy_dir = os.path.join(save_root_dir,str(case_id),str(image_num))
+            
+            
+            single_image_filenames = os.listdir(single_image_dir)
+            if len(os.listdir(single_image_dir)) == 1:
+                # 2D image
+                image_file = os.path.join(single_image_dir, single_image_filenames[0])
+                save_image_array = cv2.imread(image_file) # w,h,c
+            else:
+                save_image_array = get_image(single_image_dir,single_image_filenames)
+            if not os.path.exists(save_npy_dir):
+                    os.makedirs(save_npy_dir)
+            # print(save_image_array.shape)
+            if save_image_array is not None:
+                if len(save_image_array.shape) <=  5 and len(save_image_array.shape) >=2:
+                    save_nii_path = os.path.join(save_npy_dir,single_image_name+'.nii.gz')
+                    out = sitk.GetImageFromArray(save_image_array)
+                    sitk.WriteImage(out, save_nii_path)
+                    save_image_series.append(save_nii_path)
+                else:
+                    save_npy_path = os.path.join(save_npy_dir,single_image_name+'.npy')
+                    np.save(save_npy_path,save_image_array)
+                    save_image_series.append(save_npy_path)
+        case_dict['image'] = save_image_series
+        case_dict['image_caption'] = image_caption
+        case_dict['image_modality'] = image_modality
+        save_case_dict.append(case_dict)
+    
+if __name__ == "__main__":
+    # case_id,image_root_dir,json_root_dir
+    import argparse
+    parser = argparse.ArgumentParser()
+    parser.add_argument('--index', default=0, type=int)
+    parser.add_argument('--add_index', default=0, type=int)
+    parser.add_argument('--start_index', default=1, type=int)
+    parser.add_argument('--end_index', default=1000, type=int)
+    args = parser.parse_args()
+    
+    image_root_dir = '/mnt/petrelfs/share_data/zhangxiaoman/DATA/Radio_VQA/processed_file/images'
+    json_root_dir = '/mnt/petrelfs/share_data/zhangxiaoman/DATA/Radio_VQA/processed_file/jsons'
+    save_root_dir = '/mnt/petrelfs/share_data/zhangxiaoman/DATA/Radio_VQA/processed_file/npys'
+    save_case_dict = []
+    
+    args.start_index = args.index*1000+1 + args.add_index
+    args.end_index = (args.index+1)*1000+1
+    
+    for case_id in tqdm(range(args.start_index,args.end_index)):
+        case_id = str(case_id)
+        convert_case(case_id,image_root_dir,json_root_dir,save_case_dict,save_root_dir)
+        # CT_0 (200, 630, 630, 3)
+
+    # save to csv
+    save_json_file = '/mnt/petrelfs/share_data/zhangxiaoman/DATA/Radio_VQA/processed_file/processed_jsons/processed_json_'+str(args.index)+'.json'
+    with open(save_json_file, 'w', encoding='utf-8') as f:
+        json.dump(save_case_dict, f, ensure_ascii=False,indent=4)
+    # B, S, T, W, H, Z
+    # srun --partition=medai --mpi=pmi2 --quotatype=auto --gres=gpu:0 -n1 --ntasks-per-node=1  python data_convert.py --index 2 --add_index 24
+    # cd /mnt/petrelfs/share_data/zhangxiaoman/DATA/Radio_VQA/jpeg2npy
\ No newline at end of file