Initial commit

backend/.gitignore

@@ -1,6 +1,7 @@
 venv/
 venv310/
-.venv/
 __pycache__/
+*.pyc
 .ipynb_checkpoints/
-*.ipynb
+*.ipynb
+temp/

backend/app.py

@@ -119,10 +119,10 @@ def predict():
     except Exception as e:
 
         print("\n❌ ERROR:")
-        print(e)
+        print(str(e))
 
         return jsonify({
-            "error": "Gagal melakukan prediksi"
+            "error": str(e)
         }), 500
 
     finally:

backend/augmentasi.py

@@ -0,0 +1,108 @@
+import os
+import numpy as np
+from tensorflow.keras.preprocessing import image
+from tensorflow.keras.preprocessing.image import ImageDataGenerator
+from PIL import Image
+
+# =========================
+# PATH GAMBAR ASLI
+# =========================
+img_path = r"D:\Corn_Detection\dataset_1150\Hawar\Hawar_0118.jpg"
+
+# =========================
+# FOLDER OUTPUT
+# =========================
+output_dir = r"D:\Corn_Detection\augmentation"
+os.makedirs(output_dir, exist_ok=True)
+
+# =========================
+# LOAD IMAGE
+# =========================
+img = image.load_img(img_path, target_size=(224,224))
+img_array = image.img_to_array(img)
+img_array = np.expand_dims(img_array, axis=0)
+
+# Simpan gambar asli
+img.save(os.path.join(output_dir, "asli.jpg"))
+
+# =========================
+# FUNGSI SIMPAN AUGMENTASI
+# =========================
+def save_augmented_image(datagen, filename):
+    aug_iter = datagen.flow(img_array, batch_size=1)
+
+    batch = next(aug_iter)
+
+    img_aug = batch[0].astype("uint8")
+
+    Image.fromarray(img_aug).save(
+        os.path.join(output_dir, filename)
+    )
+
+# =========================
+# ROTATION
+# =========================
+rotation_gen = ImageDataGenerator(
+    rotation_range=20
+)
+
+save_augmented_image(rotation_gen, "rotasi.jpg")
+
+# =========================
+# ZOOM
+# =========================
+zoom_gen = ImageDataGenerator(
+    zoom_range=0.2
+)
+
+save_augmented_image(zoom_gen, "zoom.jpg")
+
+# =========================
+# HORIZONTAL FLIP
+# =========================
+flip_gen = ImageDataGenerator(
+    horizontal_flip=True
+)
+
+save_augmented_image(flip_gen, "horizontal_flip.jpg")
+
+# =========================
+# BRIGHTNESS
+# =========================
+brightness_gen = ImageDataGenerator(
+    brightness_range=[0.8,1.2]
+)
+
+save_augmented_image(brightness_gen, "brightness.jpg")
+
+# =========================
+# WIDTH SHIFT
+# =========================
+width_shift_gen = ImageDataGenerator(
+    width_shift_range=0.1
+)
+
+save_augmented_image(width_shift_gen, "width_shift.jpg")
+
+# =========================
+# HEIGHT SHIFT
+# =========================
+height_shift_gen = ImageDataGenerator(
+    height_shift_range=0.1
+)
+
+save_augmented_image(height_shift_gen, "height_shift.jpg")
+
+# =========================
+# SHEAR
+# =========================
+shear_gen = ImageDataGenerator(
+    shear_range=0.1
+)
+
+save_augmented_image(shear_gen, "shear.jpg")
+
+print("===================================")
+print("Semua hasil augmentasi berhasil disimpan")
+print("Lokasi :", output_dir)
+print("===================================")

backend/model/classification_report.csv

@@ -0,0 +1,8 @@
+,precision,recall,f1-score,support
+Bukan Jagung,0.9956521739130435,0.9956521739130435,0.9956521739130435,230.0
+Hawar,0.9912663755458515,0.9869565217391304,0.9891067538126361,230.0
+Karat,0.991304347826087,0.991304347826087,0.991304347826087,230.0
+Sehat,0.9956709956709957,1.0,0.9978308026030369,230.0
+accuracy,0.9934782608695653,0.9934782608695653,0.9934782608695653,0.9934782608695653
+macro avg,0.9934734732389945,0.9934782608695653,0.9934735195387009,920.0
+weighted avg,0.9934734732389945,0.9934782608695653,0.9934735195387009,920.0

backend/model/saved_model/fingerprint.pb

@@ -0,0 +1 @@
+û¯³‘¯‰ÔÆ•
㳜÷ÍÚ♫
À†Ë‘à—ê<E28094>u ·±™»ÝúÌŠc(¤ŽÜØš©‚Ã×
2
:&64332632573585195873750373575487188261

backend/model/training_log_fold_1.csv

@@ -0,0 +1,21 @@
+epoch,accuracy,learning_rate,loss,val_accuracy,val_loss
+0,0.7136548757553101,9.999999747378752e-05,0.7565154433250427,0.95652174949646,0.19574424624443054
+1,0.9290081262588501,9.999999747378752e-05,0.23914459347724915,0.976902186870575,0.1129738986492157
+2,0.9398776888847351,9.999999747378752e-05,0.1800134927034378,0.9755434989929199,0.08709654957056046
+3,0.9568614363670349,9.999999747378752e-05,0.13939610123634338,0.97826087474823,0.08379809558391571
+4,0.9616168737411499,9.999999747378752e-05,0.11394604295492172,0.9823369383811951,0.06505417078733444
+5,0.97044837474823,9.999999747378752e-05,0.09785516560077667,0.9836956262588501,0.058405403047800064
+6,0.9714673757553101,9.999999747378752e-05,0.0900689959526062,0.9864130616188049,0.05449285730719566
+7,0.9714673757553101,9.999999747378752e-05,0.08338520675897598,0.98777174949646,0.04825025424361229
+8,0.981317937374115,9.999999747378752e-05,0.0648878887295723,0.9864130616188049,0.04614263400435448
+9,0.977921187877655,9.999999747378752e-05,0.06618771702051163,0.9864130616188049,0.043803561478853226
+10,0.976902186870575,9.999999747378752e-05,0.06501197069883347,0.98777174949646,0.04192778468132019
+11,0.9786005616188049,9.999999747378752e-05,0.06083962693810463,0.98777174949646,0.043432462960481644
+12,0.9816576242446899,9.999999747378752e-05,0.0514533556997776,0.98777174949646,0.037732407450675964
+13,0.9884510636329651,9.999999747378752e-05,0.04559719190001488,0.98777174949646,0.03609374910593033
+14,0.984714686870575,9.999999747378752e-05,0.046360716223716736,0.99048912525177,0.036957431584596634
+15,0.9864130616188049,9.999999747378752e-05,0.04609496518969536,0.9850543737411499,0.04189329221844673
+16,0.9867526888847351,4.999999873689376e-05,0.04231247678399086,0.98777174949646,0.03806684538722038
+17,0.98607337474823,4.999999873689376e-05,0.040264297276735306,0.98777174949646,0.03658584505319595
+18,0.98607337474823,2.499999936844688e-05,0.046675294637680054,0.99048912525177,0.03497152402997017
+19,0.98607337474823,2.499999936844688e-05,0.04432199150323868,0.989130437374115,0.03393092378973961

backend/model/training_log_fold_2.csv

@@ -0,0 +1,20 @@
+epoch,accuracy,learning_rate,loss,val_accuracy,val_loss
+0,0.7381114363670349,9.999999747378752e-05,0.7068241834640503,0.9375,0.23247376084327698
+1,0.930027186870575,9.999999747378752e-05,0.24283264577388763,0.9551630616188049,0.13895472884178162
+2,0.9429348111152649,9.999999747378752e-05,0.1795891523361206,0.9646739363670349,0.09673672169446945
+3,0.9609375,9.999999747378752e-05,0.12984155118465424,0.970108687877655,0.08572567254304886
+4,0.9599184989929199,9.999999747378752e-05,0.11610609292984009,0.97826087474823,0.06869582086801529
+5,0.963994562625885,9.999999747378752e-05,0.10104407370090485,0.98097825050354,0.05855090171098709
+6,0.9745244383811951,9.999999747378752e-05,0.0835578665137291,0.9850543737411499,0.05827253311872482
+7,0.9765625,9.999999747378752e-05,0.0785723403096199,0.9823369383811951,0.05227142199873924
+8,0.977921187877655,9.999999747378752e-05,0.06692307442426682,0.9823369383811951,0.056870054453611374
+9,0.981317937374115,9.999999747378752e-05,0.06325101107358932,0.98777174949646,0.0430879108607769
+10,0.976222813129425,9.999999747378752e-05,0.06722191721200943,0.98777174949646,0.0400400385260582
+11,0.97826087474823,9.999999747378752e-05,0.056177493184804916,0.99048912525177,0.03674279898405075
+12,0.98097825050354,9.999999747378752e-05,0.05626058578491211,0.98777174949646,0.039747752249240875
+13,0.983016312122345,9.999999747378752e-05,0.04855966567993164,0.9932065010070801,0.03457866981625557
+14,0.984035313129425,9.999999747378752e-05,0.049412332475185394,0.989130437374115,0.03813661262392998
+15,0.9853940010070801,9.999999747378752e-05,0.04102384299039841,0.989130437374115,0.03746185451745987
+16,0.987432062625885,4.999999873689376e-05,0.04037996754050255,0.99048912525177,0.03834734484553337
+17,0.9881114363670349,4.999999873689376e-05,0.04008108749985695,0.991847813129425,0.03469296917319298
+18,0.98879075050354,2.499999936844688e-05,0.03981376439332962,0.9932065010070801,0.03467719256877899

backend/model/training_log_fold_3.csv

@@ -0,0 +1,21 @@
+epoch,accuracy,learning_rate,loss,val_accuracy,val_loss
+0,0.734035313129425,9.999999747378752e-05,0.6999317407608032,0.9442934989929199,0.2147262841463089
+1,0.9273098111152649,9.999999747378752e-05,0.24060937762260437,0.9646739363670349,0.13419541716575623
+2,0.94701087474823,9.999999747378752e-05,0.17050962150096893,0.9728260636329651,0.09980986267328262
+3,0.9572010636329651,9.999999747378752e-05,0.1342485100030899,0.97826087474823,0.07800345122814178
+4,0.9660326242446899,9.999999747378752e-05,0.10844268649816513,0.98097825050354,0.06890060752630234
+5,0.9667119383811951,9.999999747378752e-05,0.10096748918294907,0.98097825050354,0.0635373666882515
+6,0.971807062625885,9.999999747378752e-05,0.0858895480632782,0.9864130616188049,0.05018732696771622
+7,0.9765625,9.999999747378752e-05,0.07714884728193283,0.9836956262588501,0.05103155970573425
+8,0.973505437374115,9.999999747378752e-05,0.07740459591150284,0.9823369383811951,0.053474295884370804
+9,0.9806385636329651,4.999999873689376e-05,0.06630785763263702,0.9864130616188049,0.045486435294151306
+10,0.9806385636329651,4.999999873689376e-05,0.058046672493219376,0.98777174949646,0.04103342071175575
+11,0.9823369383811951,4.999999873689376e-05,0.055304937064647675,0.989130437374115,0.041949912905693054
+12,0.9836956262588501,4.999999873689376e-05,0.05309256166219711,0.989130437374115,0.03986254334449768
+13,0.9819973111152649,4.999999873689376e-05,0.05787506327033043,0.989130437374115,0.037597913295030594
+14,0.98267662525177,4.999999873689376e-05,0.05659506097435951,0.991847813129425,0.03598109260201454
+15,0.983016312122345,4.999999873689376e-05,0.05229465290904045,0.99048912525177,0.03660966083407402
+16,0.9819973111152649,4.999999873689376e-05,0.05457111448049545,0.99048912525177,0.03671116754412651
+17,0.98607337474823,2.499999936844688e-05,0.04707670584321022,0.98777174949646,0.03703512251377106
+18,0.98879075050354,2.499999936844688e-05,0.04428504779934883,0.989130437374115,0.03504762053489685
+19,0.9867526888847351,2.499999936844688e-05,0.04896799474954605,0.989130437374115,0.03598404303193092

backend/model/training_log_fold_4.csv

@@ -0,0 +1,21 @@
+epoch,accuracy,learning_rate,loss,val_accuracy,val_loss
+0,0.748641312122345,9.999999747378752e-05,0.6748814582824707,0.9510869383811951,0.1968584656715393
+1,0.926630437374115,9.999999747378752e-05,0.23904117941856384,0.9660326242446899,0.12440376728773117
+2,0.94972825050354,9.999999747378752e-05,0.1546500027179718,0.976902186870575,0.1011323630809784
+3,0.960597813129425,9.999999747378752e-05,0.13199861347675323,0.9850543737411499,0.07602649927139282
+4,0.9633151888847351,9.999999747378752e-05,0.11540684849023819,0.9741848111152649,0.07963371276855469
+5,0.9677309989929199,9.999999747378752e-05,0.09640498459339142,0.9836956262588501,0.059806425124406815
+6,0.9741848111152649,9.999999747378752e-05,0.08353486657142639,0.98097825050354,0.06453259289264679
+7,0.9765625,9.999999747378752e-05,0.07739967107772827,0.979619562625885,0.0524170957505703
+8,0.9786005616188049,9.999999747378752e-05,0.07381828129291534,0.9850543737411499,0.049863629043102264
+9,0.979619562625885,9.999999747378752e-05,0.06559862196445465,0.9864130616188049,0.04981501027941704
+10,0.9786005616188049,9.999999747378752e-05,0.06449061632156372,0.98097825050354,0.05683763697743416
+11,0.98267662525177,4.999999873689376e-05,0.055478502064943314,0.9850543737411499,0.049218062311410904
+12,0.98267662525177,4.999999873689376e-05,0.05234510824084282,0.9864130616188049,0.04511260613799095
+13,0.984375,4.999999873689376e-05,0.05012645572423935,0.98777174949646,0.04537223279476166
+14,0.9850543737411499,4.999999873689376e-05,0.04811901971697807,0.9864130616188049,0.04349565878510475
+15,0.985733687877655,4.999999873689376e-05,0.04842999577522278,0.9864130616188049,0.040977030992507935
+16,0.9867526888847351,4.999999873689376e-05,0.047277845442295074,0.9864130616188049,0.039208367466926575
+17,0.985733687877655,4.999999873689376e-05,0.042227428406476974,0.9864130616188049,0.039824478328228
+18,0.98777174949646,4.999999873689376e-05,0.039644014090299606,0.9850543737411499,0.03922777995467186
+19,0.9850543737411499,2.499999936844688e-05,0.04659169539809227,0.9864130616188049,0.041244834661483765

backend/model/training_log_fold_5.csv

@@ -0,0 +1,21 @@
+epoch,accuracy,learning_rate,loss,val_accuracy,val_loss
+0,0.7214673757553101,9.999999747378752e-05,0.7529850006103516,0.94701087474823,0.21970611810684204
+1,0.9191576242446899,9.999999747378752e-05,0.25084418058395386,0.95923912525177,0.12908491492271423
+2,0.950067937374115,9.999999747378752e-05,0.16257192194461823,0.970108687877655,0.09431818127632141
+3,0.9582201242446899,9.999999747378752e-05,0.13206881284713745,0.976902186870575,0.07742702215909958
+4,0.9677309989929199,9.999999747378752e-05,0.10626088827848434,0.9741848111152649,0.06937189400196075
+5,0.969089686870575,9.999999747378752e-05,0.09285038709640503,0.97826087474823,0.06072676554322243
+6,0.9667119383811951,9.999999747378752e-05,0.09316759556531906,0.976902186870575,0.05902751162648201
+7,0.9724864363670349,9.999999747378752e-05,0.07857637107372284,0.97826087474823,0.05389025807380676
+8,0.976902186870575,9.999999747378752e-05,0.07135744392871857,0.98097825050354,0.05123810097575188
+9,0.98267662525177,9.999999747378752e-05,0.06045383960008621,0.9823369383811951,0.04801933839917183
+10,0.97995924949646,9.999999747378752e-05,0.06303825229406357,0.9836956262588501,0.04756370931863785
+11,0.97826087474823,9.999999747378752e-05,0.06300406157970428,0.9823369383811951,0.04598434641957283
+12,0.9806385636329651,9.999999747378752e-05,0.05282346531748772,0.9823369383811951,0.0442848727107048
+13,0.9816576242446899,9.999999747378752e-05,0.05177825316786766,0.9836956262588501,0.044551920145750046
+14,0.984714686870575,9.999999747378752e-05,0.04417676851153374,0.98097825050354,0.048150435090065
+15,0.984375,4.999999873689376e-05,0.047107867896556854,0.9850543737411499,0.04266688972711563
+16,0.9870923757553101,4.999999873689376e-05,0.03834306448698044,0.9850543737411499,0.04022631794214249
+17,0.9898098111152649,4.999999873689376e-05,0.03330085054039955,0.9850543737411499,0.041328392922878265
+18,0.9870923757553101,4.999999873689376e-05,0.040723804384469986,0.9850543737411499,0.03947565332055092
+19,0.9898098111152649,4.999999873689376e-05,0.03469662740826607,0.9836956262588501,0.0406419076025486

backend/resize.py

@@ -0,0 +1,16 @@
+import os
+from tensorflow.keras.preprocessing import image
+
+input_path = r"D:\Corn_Detection\dataset_1150\Hawar\Hawar_0118.jpg"
+
+output_folder = r"D:\Corn_Detection\preprocessing_output\Hawar"
+
+os.makedirs(output_folder, exist_ok=True)
+
+img = image.load_img(input_path, target_size=(224,224))
+
+output_path = os.path.join(output_folder, "resize_Hawar_0118.jpg")
+
+img.save(output_path)
+
+print("Berhasil disimpan:", output_path)

backend/segmentasi.py

@@ -0,0 +1,204 @@
+import cv2
+import numpy as np
+import os
+
+# =========================
+# PATH INPUT
+# =========================
+img_path = r"D:\Corn_Detection\dataset_1150\Hawar\Hawar_0118.jpg"
+
+# =========================
+# PATH OUTPUT
+# =========================
+OUTPUT_DIR = r"D:\Corn_Detection\Segmentasi\Hawar"
+
+os.makedirs(OUTPUT_DIR, exist_ok=True)
+
+# =========================
+# BACA GAMBAR
+# =========================
+img = cv2.imread(img_path)
+
+if img is None:
+    print("Gambar tidak ditemukan!")
+    exit()
+
+img_rgb = cv2.cvtColor(
+    img,
+    cv2.COLOR_BGR2RGB
+)
+
+# =========================
+# KONVERSI HSV
+# =========================
+hsv = cv2.cvtColor(
+    img,
+    cv2.COLOR_BGR2HSV
+)
+
+# =========================
+# RENTANG WARNA DAUN
+# =========================
+lower_green = np.array([15, 20, 20])
+upper_green = np.array([110, 255, 255])
+
+mask = cv2.inRange(
+    hsv,
+    lower_green,
+    upper_green
+)
+
+# =========================
+# MORPHOLOGY
+# =========================
+kernel = np.ones((7,7), np.uint8)
+
+mask = cv2.morphologyEx(
+    mask,
+    cv2.MORPH_CLOSE,
+    kernel,
+    iterations=2
+)
+
+mask = cv2.dilate(
+    mask,
+    kernel,
+    iterations=2
+)
+
+# =========================
+# CARI KONTUR TERBESAR
+# =========================
+contours, _ = cv2.findContours(
+    mask,
+    cv2.RETR_EXTERNAL,
+    cv2.CHAIN_APPROX_SIMPLE
+)
+
+if len(contours) == 0:
+    print("Kontur daun tidak ditemukan")
+    exit()
+
+largest = max(
+    contours,
+    key=cv2.contourArea
+)
+
+# =========================
+# BOUNDING BOX
+# =========================
+x, y, w, h = cv2.boundingRect(
+    largest
+)
+
+# Tambah margin
+padding = 20
+
+x = max(0, x - padding)
+y = max(0, y - padding)
+
+w = min(
+    img.shape[1] - x,
+    w + (padding * 2)
+)
+
+h = min(
+    img.shape[0] - y,
+    h + (padding * 2)
+)
+
+# =========================
+# CROP DAUN
+# =========================
+crop_leaf = img_rgb[
+    y:y+h,
+    x:x+w
+]
+
+# =========================
+# BACKGROUND REMOVAL
+# =========================
+background_removed = cv2.bitwise_and(
+    img_rgb,
+    img_rgb,
+    mask=mask
+)
+
+# =========================
+# BOUNDING BOX IMAGE
+# =========================
+bbox_img = img_rgb.copy()
+
+cv2.rectangle(
+    bbox_img,
+    (x, y),
+    (x+w, y+h),
+    (255, 0, 0),
+    3
+)
+
+# =========================
+# MASK BERWARNA
+# =========================
+mask_color = cv2.cvtColor(
+    mask,
+    cv2.COLOR_GRAY2BGR
+)
+
+# =========================
+# SIMPAN OUTPUT
+# =========================
+cv2.imwrite(
+    os.path.join(
+        OUTPUT_DIR,
+        "0_original.jpg"
+    ),
+    img
+)
+
+cv2.imwrite(
+    os.path.join(
+        OUTPUT_DIR,
+        "1_mask.jpg"
+    ),
+    mask_color
+)
+
+cv2.imwrite(
+    os.path.join(
+        OUTPUT_DIR,
+        "2_bounding_box.jpg"
+    ),
+    cv2.cvtColor(
+        bbox_img,
+        cv2.COLOR_RGB2BGR
+    )
+)
+
+cv2.imwrite(
+    os.path.join(
+        OUTPUT_DIR,
+        "3_crop_leaf.jpg"
+    ),
+    cv2.cvtColor(
+        crop_leaf,
+        cv2.COLOR_RGB2BGR
+    )
+)
+
+cv2.imwrite(
+    os.path.join(
+        OUTPUT_DIR,
+        "4_background_removed.jpg"
+    ),
+    cv2.cvtColor(
+        background_removed,
+        cv2.COLOR_RGB2BGR
+    )
+)
+
+print("================================")
+print("Segmentasi berhasil")
+print("Output tersimpan di:")
+print(OUTPUT_DIR)
+print("================================")

backend/src/predict.py

@@ -1,13 +1,12 @@
 # predict.py
 
 import os
+import pickle
 import numpy as np
 import tensorflow as tf
 
 from tensorflow.keras.preprocessing import image
 from tensorflow.keras.applications.mobilenet_v2 import preprocess_input
-
-# KERAS 3 - LOAD SAVEDMODEL
 from keras.layers import TFSMLayer
 
 # ===============================
@@ -16,46 +15,58 @@ from keras.layers import TFSMLayer
 os.environ["TF_CPP_MIN_LOG_LEVEL"] = "2"
 
 # ===============================
-# BASE PATH
+# BASE DIR
 # ===============================
-BASE_DIR = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
+BASE_DIR = os.path.dirname(
+    os.path.dirname(
+        os.path.abspath(__file__)
+    )
+)
 
 # ===============================
 # PATH MODEL
 # ===============================
-MODEL1_PATH = os.path.join(BASE_DIR, "model_model1", "saved_model1")
-MODEL2_PATH = os.path.join(BASE_DIR, "model_model2", "saved_model2")
+MODEL_PATH = os.path.join(
+    BASE_DIR,
+    "model",
+    "saved_model"
+)
+
+CLASS_PATH = os.path.join(
+    BASE_DIR,
+    "model",
+    "class_names1.pkl"
+)
 
 # ===============================
 # LOAD MODEL
 # ===============================
-print("Loading Model 1...")
-model1 = TFSMLayer(MODEL1_PATH, call_endpoint="serve")
+print("Loading Model...")
 
-print("Loading Model 2...")
-model2 = TFSMLayer(MODEL2_PATH, call_endpoint="serve")
+model = TFSMLayer(
+    MODEL_PATH,
+    call_endpoint="serve"
+)
 
-print("Model berhasil dimuat ✔")
+print("✅ Model berhasil dimuat")
 
 # ===============================
-# LABEL
+# LOAD CLASS NAMES
 # ===============================
-LABELS_MODEL1 = ["jagung", "non_jagung"]
+with open(CLASS_PATH, "rb") as f:
+    CLASS_NAMES = pickle.load(f)
 
-LABELS_MODEL2 = [
-    "hawar_daun",
-    "karat_daun",
-    "sehat"
-]
+print("Class Names:")
+print(CLASS_NAMES)
 
 # ===============================
 # PREPROCESS IMAGE
 # ===============================
-def preprocess_image(img_path, size):
+def preprocess_image(img_path):
 
     img = image.load_img(
         img_path,
-        target_size=size
+        target_size=(224, 224)
     )
 
     img_array = image.img_to_array(img)
@@ -65,7 +76,9 @@ def preprocess_image(img_path, size):
         axis=0
     )
 
-    img_array = preprocess_input(img_array)
+    img_array = preprocess_input(
+        img_array
+    )
 
     return img_array
 
@@ -74,77 +87,42 @@ def preprocess_image(img_path, size):
 # ===============================
 def predict_image(image_path):
 
-    # ===========================
-    # CHECK FILE
-    # ===========================
     if not os.path.exists(image_path):
+
         raise FileNotFoundError(
             "File gambar tidak ditemukan"
         )
 
-    print("\n==============================")
-    print("PREDIKSI GAMBAR")
-    print("==============================")
-
-    # ===========================
-    # MODEL 1
-    # JAGUNG / NON JAGUNG
-    # ===========================
-    img1 = preprocess_image(
-        image_path,
-        (128, 128)
+    img = preprocess_image(
+        image_path
     )
 
-    preds1 = model1(img1)
-    preds1 = preds1.numpy()[0]
+    preds = model(img)
 
-    idx1 = np.argmax(preds1)
+    preds = preds.numpy()[0]
 
-    label1 = LABELS_MODEL1[idx1]
+    idx = np.argmax(preds)
 
-    conf1 = float(preds1[idx1]) * 100
+    label = CLASS_NAMES[idx]
 
-    print("\n--- MODEL 1 ---")
-    print("Prediksi :", label1)
-    print("Confidence :", f"{conf1:.2f}%")
-
-    # ===========================
-    # JIKA NON JAGUNG
-    # ===========================
-    if label1 == "non_jagung":
-
-        return {
-            "status": "non_jagung",
-            "penyakit": None,
-            "confidence": round(conf1, 2)
-        }
-
-    # ===========================
-    # MODEL 2
-    # PENYAKIT JAGUNG
-    # ===========================
-    img2 = preprocess_image(
-        image_path,
-        (224, 224)
-    )
-
-    preds2 = model2(img2)
-    preds2 = preds2.numpy()[0]
-
-    idx2 = np.argmax(preds2)
-
-    label2 = LABELS_MODEL2[idx2]
-
-    conf2 = float(preds2[idx2]) * 100
-
-    print("\n--- MODEL 2 ---")
-    print("Prediksi :", label2)
-    print("Confidence :", f"{conf2:.2f}%")
+    confidence = float(
+        preds[idx]
+    ) * 100
 
     return {
-        "status": "jagung",
-        "penyakit": label2,
-        "confidence": round(conf2, 2)
+        "kelas": label,
+        "confidence": round(
+            confidence,
+            2
+        ),
+        "probabilities": {
+            CLASS_NAMES[i]:
+            round(
+                float(preds[i]) * 100,
+                2
+            )
+            for i in range(len(CLASS_NAMES))
+        }
     }
 
 # ===============================
@@ -154,13 +132,29 @@ if __name__ == "__main__":
 
     IMAGE_PATH = os.path.join(
         BASE_DIR,
-        "test_image5.jpg"
+        "test_image4.jpg"
     )
 
-    result = predict_image(IMAGE_PATH)
+    result = predict_image(
+        IMAGE_PATH
+    )
 
-    print("\n==============================")
-    print("HASIL AKHIR")
-    print("==============================")
+    print("\n====================")
+    print("HASIL PREDIKSI")
+    print("====================")
 
-    print(result)
+    print(
+        f"Kelas      : {result['kelas']}"
+    )
+
+    print(
+        f"Confidence : {result['confidence']}%"
+    )
+
+    print("\nProbabilitas:")
+
+    for k, v in result["probabilities"].items():
+
+        print(
+            f"{k:<15}: {v:.2f}%"
+        )

backend/static/js/script.js

@@ -46,7 +46,6 @@ document.addEventListener("DOMContentLoaded", function () {
 
     if (!file) return;
 
-    // Validasi gambar
     if (!file.type.startsWith("image/")) {
       alert("File harus berupa gambar!");
       input.value = "";
@@ -55,13 +54,11 @@ document.addEventListener("DOMContentLoaded", function () {
 
     selectedFile = file;
 
-    // Preview gambar
     if (previewImage) {
       previewImage.src = URL.createObjectURL(file);
       previewImage.style.display = "block";
     }
 
-    // Reset hasil lama
     if (resultBox) {
       resultBox.style.display = "none";
     }
@@ -99,7 +96,6 @@ document.addEventListener("DOMContentLoaded", function () {
     const formData = new FormData();
     formData.append("image", selectedFile);
 
-    // Tampilkan loading
     if (resultBox) {
       resultBox.style.display = "block";
     }
@@ -121,9 +117,13 @@ document.addEventListener("DOMContentLoaded", function () {
 
       const data = await response.json();
 
-      console.log(data);
+      console.log("Hasil Prediksi:", data);
 
-      // Error backend
+      /**
+       * ==========================================
+       * ERROR BACKEND
+       * ==========================================
+       */
       if (!response.ok) {
 
         if (diseaseName) {
@@ -140,49 +140,39 @@ document.addEventListener("DOMContentLoaded", function () {
 
       /**
        * ==========================================
-       * HASIL NON JAGUNG
+       * HASIL PREDIKSI
        * ==========================================
        */
-      if (data.status === "non_jagung") {
+      let kelas = data.kelas;
 
-        if (diseaseName) {
-          diseaseName.textContent =
-            "❌ Gambar bukan daun jagung";
-        }
-
-        if (confidenceText) {
-          confidenceText.textContent =
-            `Tingkat Kepercayaan: ${data.confidence}%`;
-        }
-
-        return;
-      }
-
-      /**
-       * ==========================================
-       * HASIL JAGUNG
-       * ==========================================
-       */
-      if (diseaseName) {
-
-        let penyakit = data.penyakit;
-
-        // Format nama penyakit
-        if (penyakit === "hawar_daun") {
-          penyakit = "Hawar Daun";
-        }
-        else if (penyakit === "karat_daun") {
-          penyakit = "Karat Daun";
-        }
-        else if (penyakit === "sehat") {
-          penyakit = "Daun Sehat";
-        }
+      if (kelas === "Bukan Jagung") {
 
         diseaseName.textContent =
-          `🌽 Penyakit: ${penyakit}`;
+          "❌ Gambar Bukan Daun Jagung";
+
+      } else if (kelas === "Hawar") {
+
+        diseaseName.textContent =
+          "🌽 Penyakit: Hawar Daun";
+
+      } else if (kelas === "Karat") {
+
+        diseaseName.textContent =
+          "🌽 Penyakit: Karat Daun";
+
+      } else if (kelas === "Sehat") {
+
+        diseaseName.textContent =
+          "🌿 Daun Jagung Sehat";
+
+      } else {
+
+        diseaseName.textContent =
+          `🌽 Hasil Deteksi: ${kelas}`;
       }
 
       if (confidenceText) {
+
         confidenceText.textContent =
           `Tingkat Kepercayaan: ${data.confidence}%`;
       }