Update app.py

app.py

@@ -10,6 +10,7 @@ from tensorflow.keras.callbacks import EarlyStopping
 from sklearn.utils.class_weight import compute_class_weight
 from sklearn.model_selection import train_test_split
 from tensorflow.image import resize
+from google.colab.patches import cv2_imshow
 
 from tensorflow.keras.utils import to_categorical
 import matplotlib.pyplot as plt
@@ -39,106 +40,68 @@ models = {
 model_name = st.selectbox("Choose a model", list(models.keys()))
 model = models[model_name]
 
-# Upload CSV file 
-# file = st.file_uploader("Upload a CSV file", type=["csv"])
-file ='hmnist_28_28_RGB.csv'
-
-def image_resize(data):
-    Data = data.drop(columns=["label"])
-    Data = np.array(Data).reshape(-1, 28, 28, 3)
-    Data = Data / 255.0  # Normalizing the data
-    # Resize images to 32x32 pixels
-    Data_resized = resize(Data, [32, 32]).numpy()  # Ensure conversion to NumPy array
-    return Data_resized
-
-if file is not None:
-    df = pd.read_csv(file)  
-    # Get first row 
-    img_reshaped = image_resize(df)
+# Upload Image
+
+file = st.file_uploader("Upload an image", type=["png", "jpg", "jpeg"])
+# file ='hmnist_28_28_RGB.csv'
+
+classes = {4: ('nv', ' melanocytic nevi'), 6: ('mel', 'melanoma'),
+           2 :('bkl', 'benign keratosis-like lesions'), 1:('bcc' , ' basal cell carcinoma'),
+           5: ('vasc', ' pyogenic granulomas and hemorrhage'),
+           0: ('akiec', 'Actinic keratoses and intraepithelial carcinomae'),
+           3: ('df', 'dermatofibroma')}
+
+
+if file is not None:   
+    img = cv2.imread(file)
+    # cv2_imshow(img)
+    img1 = cv2.resize(img, (28, 28))
+    result = model.predict(img1.reshape(1, 28, 28, 3))
+    max_prob = max(result[0])
+    class_ind = list(result[0]).index(max_prob)
+    class_name = classes[class_ind]
+    # print(class_name)
+    # count+=1
+    # if count>10:
+    #     break
+
+        
+    # df = pd.read_csv(file)  
+    # # Get first row 
+    # img_reshaped = image_resize(df)
     
-    # Get prediction
-    pred = model.predict(img_reshaped)
-    label = np.argmax(pred)
-
-    # label_map = {4: ('nv', ' melanocytic nevi'), 
-    #          6: ('mel', 'melanoma'),
-    #          2: ('bkl', 'benign keratosis-like lesions'),
-    #          1: ('bcc' , ' basal cell carcinoma'),
-    #          5: ('vasc', 'pyogenic granulomas and hemorrhage'),  
-    #          0: ('akiec', 'Actinic keratoses and intraepithelial carcinomae'),
-    #          3: ('df', 'dermatofibroma')}
-
-    # if label in label_map:
-    #     label_name = label_map[label][0]
-    #     full_name = label_map[label][1]
+    # # Get prediction
+    # pred = model.predict(img_reshaped)
+    # label = np.argmax(pred)
+
     
     # Display image and result 
     col1, col2 = st.columns(2)
     with col1:
         st.header("Input Image")
-        # st.image(image)
+        st.image(img)
     with col2:
         st.header("Prediction")
-        st.write(label)
-        # st.metric("Digit", full_name)
-
-
-# import streamlit as st
-# import predict_model # our prediction model
-
-# # Label maps
-# label_map = {0: ('akiec', 'Actinic keratoses'), 
-#              1: ('bcc', 'basal cell carcinoma'),
-#              # Rest of label map
-#              }
-
-# # Get prediction        
-# img = st.file_uploader("Upload image")
-# if img:
-#     pred_id = predict_model.get_prediction(img) 
-    
-#     # Display prediction    
-#     if pred_id in label_map:
-#         label_name = label_map[pred_id][0]
-#         full_name = label_map[pred_id][1]
-        
-#         st.success(f"Predicted Label: {label_name} - {full_name}")
-#     else:
-#         st.warning("Unknown label predicted")        
-
-        
-
-# data_dir = 'hmnist_28_28_RGB.csv'
-# data = pd.read_csv(data_dir)
-# data.head()
-
-# Label = data["label"]
-# Data = data.drop(columns=["label"])
-
-# data["label"].value_counts()
-
-# classes = {4: ('nv', ' melanocytic nevi'),
-#            6: ('mel', 'melanoma'),
-#            2 :('bkl', 'benign keratosis-like lesions'), 
-#            1:('bcc' , ' basal cell carcinoma'),
-#            5: ('vasc', ' pyogenic granulomas and hemorrhage'),
-#            0: ('akiec', 'Actinic keratoses and intraepithelial carcinomae'),
-#            3: ('df', 'dermatofibroma')}
-
-# from tensorflow.image import resize
-
-# #preprocess data
-# Label = data["label"]
-
-
-
-# Label = to_categorical(Label, num_classes=7)  # Assuming 7 classes
-
-
-
+        st.write(class_name)
+        st.metric("Category:", class_name)
 
-# # Later in Streamlit...
 
 
 
 
+# from google.colab.patches import cv2_imshow
+# srcdir = '/kaggle/input/skin-cancer-mnist-ham10000/HAM10000_images_part_1'
+# count=0
+# for temp in os.listdir(srcdir):
+#     img = cv2.imread(os.path.join(srcdir, temp))
+#     cv2.imwrite(temp, img)
+#     cv2_imshow(img)
+#     img = cv2.resize(img, (28, 28))
+#     result = model.predict(img.reshape(1, 28, 28, 3))
+#     max_prob = max(result[0])
+#     class_ind = list(result[0]).index(max_prob)
+#     class_name = classes[class_ind]
+#     print(class_name)
+#     count+=1
+#     if count>10:
+#         break