Updates

DeepQuestAI · Jan 17, 2019 · fd104c3 · fd104c3
1 parent c011f34
commit fd104c3
Show file tree

Hide file tree

Showing 6 changed files with 46 additions and 10 deletions.
diff --git a/facedetection.rst b/facedetection.rst
@@ -35,7 +35,7 @@ Result ::
     male
     male
     female
-    {'predictions': [{'y_max': 303, 'gender': 'female', 'confidence': 100, 'x_min': 534, 'x_max': 629, 'y_min': 174}, {'y_max': 275, 'gender': 'male', 'confidence': 99, 'x_min': 616, 'x_max': 711, 'y_min': 146}, {'y_max': 259, 'gender': 'male', 'confidence': 98, 'x_min': 729, 'x_max': 811, 'y_min': 147}, {'y_max': 290, 'gender': 'female', 'confidence': 99, 'x_min': 471, 'x_max': 549, 'y_min': 190}], 'success': True}
+    {'success': True, 'predictions': [{'gender': 'female', 'x_min': 534, 'x_max': 629, 'y_max': 303, 'y_min': 174, 'confidence': 0.99999213}, {'gender': 'male', 'x_min': 616, 'x_max': 711, 'y_max': 275, 'y_min': 146, 'confidence': 0.6611953}, {'gender': 'male', 'x_min': 729, 'x_max': 811, 'y_max': 259, 'y_min': 147, 'confidence': 0.99884146}, {'gender': 'female', 'x_min': 471, 'x_max': 549, 'y_max': 290, 'y_min': 190, 'confidence': 0.99997365}]}
 
 We can use the coordinates returned to extract the faces from the image
 

diff --git a/index.rst b/index.rst
@@ -34,10 +34,11 @@ Below, using DeepStack we attempt to classify the scene of the above image ::
 
 **Result**  ::
 
-    {'label': 'highway', 'confidence': 63.377846, 'success': True}
+    {'label': 'highway', 'success': True, 'confidence': 0.63377845}
+
 
 You simply send in an image by POST and deepstack returns a JSON response detailing the label of 
-the image as well as the confidence of the prediction on a scale of 1 - 100.
+the image as well as the confidence of the prediction on a scale of 0 - 1.
 
 
 **Installing DeepStack**

diff --git a/objectdetection.rst b/objectdetection.rst
@@ -45,10 +45,10 @@ If using the GPU Version, run ::
 
 Result ::
 
-    dog
     person
     person
-    {'predictions': [{'x_max': 819, 'x_min': 633, 'y_min': 354, 'confidence': 99, 'label': 'dog', 'y_max': 546}, {'x_max': 601, 'x_min': 440, 'y_min': 116, 'confidence': 99, 'label': 'person', 'y_max': 516}, {'x_max': 445, 'x_min': 295, 'y_min': 84, 'confidence': 99, 'label': 'person', 'y_max': 514}], 'success': True}
+    dog
+    {'predictions': [{'x_max': 450, 'x_min': 294, 'confidence': 0.9999949, 'y_max': 515, 'y_min': 90, 'label': 'person'}, {'x_max': 602, 'x_min': 442, 'confidence': 0.99987614, 'y_max': 524, 'y_min': 120, 'label': 'person'}, {'x_max': 792, 'x_min': 648, 'confidence': 0.99875224, 'y_max': 545, 'y_min': 350, 'label': 'dog'}], 'success': True}
 
 We can use the coordinates returned to extract the objects
 
@@ -85,6 +85,22 @@ Result
 .. figure:: image2_person.jpg
     :align: center
 
+**Setting Confidence Levels**
+
+By default, all objects with a confidence of 0.4 and above are detected. However, you can adjust the confidence level using the
+*min_confidence* parameter.
+
+Example ::
+
+    import requests
+    
+    image_data = open("test-image3.jpg","rb").read()
+    
+    response = requests.post("http://localhost:80/v1/vision/detection",files={"image":image_data},data={"min_confidence":0.6}).json()
+    
+
+In the above, only objects with 60% probability will be detected. If the confidence level is too high, many objects may not be detected,
+if it is too low, it might detect objects that are not present.
 
 **CLASSES**
 

diff --git a/releasenotes.rst b/releasenotes.rst
@@ -29,6 +29,17 @@ Improvements
     * **50% Reduction in Install Size**
 
         The total install size DeepStack is now half the original size.
+
+API Changes
+------------
+
+    * **OBJECT DETECTION - MINIMUM CONFIDENCE**
+
+        By default, all objects with a confidence of 0.4 and above are detected. However, you can adjust the confidence level using the
+        *min_confidence* parameter.
+
+
+
 
 Breaking Changes
 ----------------
@@ -48,6 +59,11 @@ Breaking Changes
 
             sudo docker run -e VISION-FACE=True -e VISION-SCENE=True -e VISION-TRAFFIC=True \
             -v localstorage:/datastore -p 80:5000 deepquestai/deepstack
+
+    * **CONFIDENCE VALUES**
+
+        All endpoints now return confidence scores in range of 0 - 1. 
+        Previously, some endpoints returned confidence in range of 0 - 100
 
     * **FACE RECOGNITION API-MINIMUM CONFIDENCE**
 
@@ -88,10 +104,12 @@ Breaking Changes
 
         To ::
 
-            {'success': True, 'confidence': 73.73981, 'label': 'conference_room'}
+            {'success': True, 'confidence': 0.7373981, 'label': 'conference_room'}
 
     * **OBJECT DECTION API - SPEED MODES**
 
         Speed Modes have been deprecated.
 
+
+
 
diff --git a/scenerecognition.rst b/scenerecognition.rst
@@ -42,6 +42,6 @@ If using the GPU Version, run ::
 
 Result ::
 
-    Label: conference_room
-    {'success': True, 'confidence': 73.73981, 'label': 'conference_room'}
+    {'confidence': 0.73739815, 'label': 'conference_room', 'success': True}
+
 
diff --git a/trafficrecognition.rst b/trafficrecognition.rst
@@ -54,5 +54,6 @@ If using the GPU Version, run ::
 
 Result ::
 
-    Label: accident
-    {'success': True, 'confidence': 94.88776, 'label': 'accident'}
+    Label accident
+    {'success': True, 'label': 'accident', 'confidence': 0.9488776}
+