diff --git a/src/Features2d.cc b/src/Features2d.cc
index fa8f1cb7..4048a2be 100644
--- a/src/Features2d.cc
+++ b/src/Features2d.cc
@@ -5,6 +5,7 @@
 #include "Matrix.h"
 #include <nan.h>
 #include <stdio.h>
+#include <limits>
 
 void Features::Init(Local<Object> target) {
   Nan::HandleScope scope;
@@ -46,38 +47,47 @@ class AsyncDetectSimilarity: public Nan::AsyncWorker {
     extractor->compute(image1, keypoints1, descriptors1);
     extractor->compute(image2, keypoints2, descriptors2);
 
-    matcher->match(descriptors1, descriptors2, matches);
-
-    double max_dist = 0;
-    double min_dist = 100;
-
-    //-- Quick calculation of max and min distances between keypoints
-    for (int i = 0; i < descriptors1.rows; i++) {
-      double dist = matches[i].distance;
-      if (dist < min_dist) {
-        min_dist = dist;
-      }
-      if (dist > max_dist) {
-        max_dist = dist;
+    //fix for 'Assertion failed...' error
+    //source: http://stackoverflow.com/questions/15650371/matcher-assertions-failed-error-opencv-android
+    if (descriptors1.type() == descriptors2.type() && descriptors1.cols == descriptors2.cols) {
+      matcher->match(descriptors1, descriptors2, matches);
+
+      double max_dist = 0;
+      double min_dist = 100;
+
+      //-- Quick calculation of max and min distances between keypoints
+      for (int i = 0; i < descriptors1.rows; i++) {
+        double dist = matches[i].distance;
+        if (dist < min_dist) {
+          min_dist = dist;
+        }
+        if (dist > max_dist) {
+          max_dist = dist;
+        }
       }
-    }
 
-    //-- Draw only "good" matches (i.e. whose distance is less than 2*min_dist,
-    //-- or a small arbitary value ( 0.02 ) in the event that min_dist is very
-    //-- small)
-    //-- PS.- radiusMatch can also be used here.
-    std::vector<cv::DMatch> good_matches;
-    double good_matches_sum = 0.0;
-
-    for (int i = 0; i < descriptors1.rows; i++) {
-      double distance = matches[i].distance;
-      if (distance <= std::max(2 * min_dist, 0.02)) {
-        good_matches.push_back(matches[i]);
-        good_matches_sum += distance;
+      //-- Draw only "good" matches (i.e. whose distance is less than 2*min_dist,
+      //-- or a small arbitary value ( 0.02 ) in the event that min_dist is very
+      //-- small)
+      //-- PS.- radiusMatch can also be used here.
+      std::vector<cv::DMatch> good_matches;
+      double good_matches_sum = 0.0;
+
+      for (int i = 0; i < descriptors1.rows; i++) {
+        double distance = matches[i].distance;
+        if (distance <= std::max(2 * min_dist, 0.02)) {
+          good_matches.push_back(matches[i]);
+          good_matches_sum += distance;
+        }
       }
-    }
 
-    dissimilarity = (double) good_matches_sum / (double) good_matches.size();
+      dissimilarity = (double) good_matches_sum / (double) good_matches.size();
+
+    }
+    else {
+      dissimilarity = std::numeric_limits<double>::quiet_NaN();
+    }
+    
   }
 
   void HandleOKCallback() {
@@ -86,7 +96,9 @@ class AsyncDetectSimilarity: public Nan::AsyncWorker {
     Local<Value> argv[2];
 
     argv[0] = Nan::Null();
-    argv[1] = Nan::New<Number>(dissimilarity);
+
+    if (dissimilarity != dissimilarity) argv[1] = Nan::Null();
+    else argv[1] = Nan::New<Number>(dissimilarity);
 
     callback->Call(2, argv);
   }