Update app.py

app.py

@@ -2,43 +2,65 @@ import gradio as gr
 from roboflow import Roboflow
 import tempfile
 import os
+import cv2
+import supervision as sv
+import numpy as np
 
 # Inisialisasi Roboflow
 rf = Roboflow(api_key="Otg64Ra6wNOgDyjuhMYU")
 project = rf.workspace("alat-pelindung-diri").project("nescafe-4base")
 model = project.version(16).model
 
-# Fungsi untuk menangani input dan output gambar
+# Fungsi untuk deteksi objek dengan supervision InferenceSlicer
 def detect_objects(image):
-    # Menyimpan gambar yang diupload sebagai file sementara
+    # Simpan gambar sementara
     with tempfile.NamedTemporaryFile(delete=False, suffix=".jpg") as temp_file:
         image.save(temp_file, format="JPEG")
         temp_file_path = temp_file.name
 
-    # Lakukan prediksi pada gambar
-    predictions = model.predict(temp_file_path, confidence=50, overlap=30).json()
-    
+    # Membaca gambar dengan OpenCV
+    img = cv2.imread(temp_file_path)
+
+    # Callback function untuk model prediksi
+    def callback(image_slice: np.ndarray) -> sv.Detections:
+        # Lakukan inferensi pada setiap potongan gambar
+        predictions = model.predict(image_slice, confidence=50, overlap=30).json()
+        return sv.Detections.from_inference(predictions)
+
+    # Menggunakan InferenceSlicer
+    slicer = sv.InferenceSlicer(callback=callback)
+
+    # Proses gambar dengan slicer
+    detections = slicer(img)
+
+    # Filter deteksi yang tumpang tindih (gunakan NMM atau tanpa filter)
+    filtered_detections = detections.filter(strategy=sv.OverlapFilter.NON_MAX_MERGE, iou_threshold=0.5)
+
+    # Annotasi gambar dengan deteksi
+    annotated_image = sv.BoxAnnotator().annotate(scene=img.copy(), detections=filtered_detections)
+
+    # Simpan gambar dengan prediksi
+    output_path = "/tmp/prediction.jpg"
+    cv2.imwrite(output_path, annotated_image)
+
+    # Hapus file sementara
+    os.remove(temp_file_path)
+
     # Menghitung jumlah objek per kelas
     class_count = {}
-    for prediction in predictions['predictions']:
-        class_name = prediction['class']
+    for detection in filtered_detections:
+        class_name = detection.class_name
         if class_name in class_count:
             class_count[class_name] += 1
         else:
             class_count[class_name] = 1
 
-    # Menyusun output berupa string hasil perhitungan
+    # Hasil perhitungan objek
     result_text = "Jumlah objek per kelas:\n"
     for class_name, count in class_count.items():
         result_text += f"{class_name}: {count} objek\n"
-    
-    # Menyimpan gambar dengan prediksi
-    output_image = model.predict(temp_file_path, confidence=50, overlap=30).save("/tmp/prediction.jpg")
-    
-    # Hapus file sementara setelah prediksi
-    os.remove(temp_file_path)
-    
-    return "/tmp/prediction.jpg", result_text
+
+    return output_path, result_text
 
 # Membuat antarmuka Gradio
 iface = gr.Interface(