Adding statistics.py to compute our model score easily and with more …

…clarity

statistics.py

@@ -0,0 +1,59 @@
+def positive_image_stats(objs_found, nb_objs):
+	"""	Update our general positive image stats.
+
+		Args:
+			objs_found : int
+				Number of objects found in image.
+			nb_objs : int
+				Total number of objects in image.
+
+		Returns:
+			tuple: (true_pos, false_neg, true_neg, false_pos)
+	"""
+	true_pos = 0
+	false_neg = 0
+	false_pos = 0
+
+	#	If our match is correct, just set the true_pos value.
+	if objs_found == nb_objs:
+		true_pos += objs_found
+	else:
+
+		#	Otherwise...
+		#	If the number of objects found is superior, the excedents are false positives...
+		#	Else... the excedents are false_negatives.
+		diff = nb_objs - objs_found
+		if diff > 0:
+			true_pos += nb_objs - diff
+			false_neg += diff
+		else:
+			true_pos = nb_objs
+			false_pos += abs(diff)
+
+	return (true_pos, false_neg, 0, false_pos)
+
+def negative_image_stats(objs_found):
+	"""	Update our general negative image stats.
+
+		Args:
+			objs_found : int
+				Number of objects found in image.
+
+		Returns:
+			tuple: (true_pos, false_neg, true_neg, false_pos)
+	"""
+	true_neg = 0
+	false_pos = 0
+	if objs_found == 0:
+		true_neg = 1
+	else:
+		false_pos = objs_found
+
+	return (0, 0, true_neg, false_pos)
+
+
+if __name__ == '__main__':
+	total_positives = 40
+	total_negatives = 40
+	print (positive_image_stats(3, 1))
+	print (negative_image_stats(0))