add draw pr_result.jpg

@@ -61,18 +61,38 @@ def get_evaluate_score(true_image_path, true_label_path, predict_label_path, thr
     p = precision_score(targets, predicts)
     r = recall_score(targets, predicts)
     conf = confusion_matrix(targets, predicts)
+    bg_mask = np.ones((500, 500, 3)) * 255
+    cv2.putText(bg_mask, f'              authentic     tampered ', (20, 50), cv2.FONT_ITALIC, 0.7, (0, 0, 255), 1)
+    cv2.putText(bg_mask, f'authentic      {conf[0, 0]}           {conf[0, 1]}', (20, 80), cv2.FONT_ITALIC, 0.7,
+                (0, 0, 255), 1)
+    cv2.putText(bg_mask, f'tempered       {conf[1, 0]}           {conf[1, 1]}', (20, 110), cv2.FONT_ITALIC, 0.7,
+                (0, 0, 255), 1)
+    cv2.putText(bg_mask, f'authentic  precision:{round(conf[0, 0] / (conf[0, 0] + conf[1, 0]), 3)}', (20, 170),
+                cv2.FONT_ITALIC, 0.7, (0, 0, 255), 1)
+    cv2.putText(bg_mask, f'           recall:{round(conf[0, 0] / (conf[0, 0] + conf[0, 1]), 3)}', (20, 200),
+                cv2.FONT_ITALIC, 0.7, (0, 0, 255), 1)
+    cv2.putText(bg_mask, f'tampered  precision:{round(conf[1, 1] / (conf[0, 1] + conf[1, 1]), 3)}', (20, 230),
+                cv2.FONT_ITALIC, 0.7, (0, 0, 255), 1)
+    cv2.putText(bg_mask, f'           recall:{round(conf[1, 1] / (conf[1, 0] + conf[1, 1]), 3)}', (20, 260),
+                cv2.FONT_ITALIC, 0.7, (0, 0, 255), 1)
+    cv2.imwrite(f'pr_result.jpg', bg_mask)
     print('precison:', p)
     print('recall:', r)
     print(conf)
     print(f'                      预    测        ')
     print(f'               authentic     tampered ')
-    print(f'真  authentic \t\t{conf[0, 0]}  \t\t{conf[0,1]}')
-    print(f'实  tempered  \t\t{conf[1, 0]}  \t\t\t{conf[1,1]}')
-    print(f'authentic precision:{conf[0,0]/(conf[0,0]+conf[1,0])}\trecall:{conf[0, 0]/(conf[0, 0]+conf[0, 1])}')
-    print(f'tampered  precision:{conf[1, 1]/(conf[0, 1]+conf[1, 1])}\trecall:{conf[1, 1]/(conf[1, 0]+conf[1, 1])}')
+    print(f'真  authentic \t\t{conf[0, 0]}  \t\t{conf[0, 1]}')
+    print(f'实  tempered  \t\t{conf[1, 0]}  \t\t\t{conf[1, 1]}')
+    print(
+        f'authentic precision:{conf[0, 0] / (conf[0, 0] + conf[1, 0])}\trecall:{conf[0, 0] / (conf[0, 0] + conf[0, 1])}')
+    print(
+        f'tampered  precision:{conf[1, 1] / (conf[0, 1] + conf[1, 1])}\trecall:{conf[1, 1] / (conf[1, 0] + conf[1, 1])}')
+
+
 if __name__ == '__main__':
-    true_image_path = '/data/situ_invoice_bill_data/qfs_train_val_data/gongshang/images/val'
-    true_label_path = '/data/situ_invoice_bill_data/qfs_train_val_data/gongshang/labels/val'
-    predict_label_path = '/home/situ/qfs/invoice_tamper/09_project/project/tamper_det/runs/detect/exp4/labels'
+    true_image_path = '/data/situ_invoice_bill_data/qfs_train_val_data/test_data/only_human_ps/all/images'
+    true_label_path = '/data/situ_invoice_bill_data/qfs_train_val_data/test_data/only_human_ps/all/labels'
+    predict_label_path = '/home/situ/qfs/invoice_tamper/09_project/project/tamper_det/runs/detect/exp2/labels'
     threshold = 0.1
     get_evaluate_score(true_image_path, true_label_path, predict_label_path, threshold)
+