diff --git a/inference-engine/src/low_precision_transformations/src/layer_transformation.cpp b/inference-engine/src/low_precision_transformations/src/layer_transformation.cpp
index dbbb8206b4d..834aa6931c5 100644
--- a/inference-engine/src/low_precision_transformations/src/layer_transformation.cpp
+++ b/inference-engine/src/low_precision_transformations/src/layer_transformation.cpp
@@ -221,18 +221,20 @@ LayerTransformation::PrecisionDetails LayerTransformation::getPrecisionDetails(c
     bool hasZeroPoint = false;
     for (size_t i = 0; i < quantizationDetails.outputLowValues.size(); ++i) {
         const bool signedInterval = std::signbit(quantizationDetails.outputLowValues[i]) != std::signbit(quantizationDetails.outputHighValues[i]);
-        const bool boundaryValuesAreNotZero =
-            (std::fabs(quantizationDetails.outputLowValues[i]) >= zeroThreshold) &&
-            (std::fabs(quantizationDetails.outputHighValues[i]) >= zeroThreshold);
-        if (signedInterval && boundaryValuesAreNotZero) {
+        const bool outputLowValueIsNotZero = std::fabs(quantizationDetails.outputLowValues[i]) >= zeroThreshold;
+        if (signedInterval && outputLowValueIsNotZero) {
             // signed
             unsignedPrecision = false;
             hasNegative = true;
 
-            const float expectedRatio = quantizationDetails.levels == 256 ? asymmetricIntervalSideRatio256 : -1.f;
-            const float actualRatio = quantizationDetails.outputLowValues[i] / quantizationDetails.outputHighValues[i];
-            const float actual = std::fabs((actualRatio - expectedRatio) / std::min(actualRatio, expectedRatio));
-            if (actual > quantizationIntervalAsymmetryThreshold) {
+            if (quantizationDetails.outputHighValues[i] != 0.f) {
+                const float expectedRatio = quantizationDetails.levels == 256 ? asymmetricIntervalSideRatio256 : -1.f;
+                const float actualRatio = quantizationDetails.outputLowValues[i] / quantizationDetails.outputHighValues[i];
+                const float actual = std::fabs((actualRatio - expectedRatio) / std::min(actualRatio, expectedRatio));
+                if (actual > quantizationIntervalAsymmetryThreshold) {
+                    hasZeroPoint = true;
+                }
+            } else {
                 hasZeroPoint = true;
             }
 #ifdef LPT_PRINT_DEQUANTIZATION_INFO
@@ -244,8 +246,8 @@ LayerTransformation::PrecisionDetails LayerTransformation::getPrecisionDetails(c
         } else {
             // unsigned
             signedPrecision = false;
-            if (boundaryValuesAreNotZero) {
-                hasZeroPoint = boundaryValuesAreNotZero;
+            if (outputLowValueIsNotZero) {
+                hasZeroPoint = outputLowValueIsNotZero;
             }
 
 #ifdef LPT_PRINT_DEQUANTIZATION_INFO
diff --git a/inference-engine/tests/functional/plugin/cpu/shared_tests_instances/low_precision_transformations/fake_quantize_transformation.cpp b/inference-engine/tests/functional/plugin/cpu/shared_tests_instances/low_precision_transformations/fake_quantize_transformation.cpp
index 2f856a61cd7..0208a6b0e72 100644
--- a/inference-engine/tests/functional/plugin/cpu/shared_tests_instances/low_precision_transformations/fake_quantize_transformation.cpp
+++ b/inference-engine/tests/functional/plugin/cpu/shared_tests_instances/low_precision_transformations/fake_quantize_transformation.cpp
@@ -42,6 +42,10 @@ const std::vector<FakeQuantizeTransformationParam> fakeQuantizeOnDataValues = {
         { 256ul, {}, { 0.f }, { 2.55f }, { 2.55f }, { 2.55f } },
         "Pooling", "U8"
     },
+    {
+        { 256ul, {}, { -127.5f }, { 0.f }, { -127.5f }, { 0.f } },
+        "Pooling", "U8"
+    },
     {
         { 16ul, {}, { 0.f }, { 1.5f }, { 0.f }, { 1.5f } },
         "Pooling", "FP32"
@@ -50,7 +54,7 @@ const std::vector<FakeQuantizeTransformationParam> fakeQuantizeOnDataValues = {
         { 16ul, {}, { -8.f }, { 7.f }, { -0.8f }, { 0.7f } },
         "Pooling", "FP32"
     },
-        // nGraph: I8->FP32 Convert is not supported
+    // nGraph: I8->FP32 Convert is not supported
     // { 256ul, {}, { -1.28f} , { 1.27f }, { -1.28f} , { 1.27f } },
     // { 256ul, { 1ul }, { -1.28f} , { 1.27f } }
 };
diff --git a/inference-engine/tests/functional/plugin/gpu/shared_tests_instances/low_precision_transformations/fake_quantize_transformation.cpp b/inference-engine/tests/functional/plugin/gpu/shared_tests_instances/low_precision_transformations/fake_quantize_transformation.cpp
index 35f047794da..62fb144c785 100644
--- a/inference-engine/tests/functional/plugin/gpu/shared_tests_instances/low_precision_transformations/fake_quantize_transformation.cpp
+++ b/inference-engine/tests/functional/plugin/gpu/shared_tests_instances/low_precision_transformations/fake_quantize_transformation.cpp
@@ -26,30 +26,34 @@ const std::vector<LayerTransformation::Params> trasformationParamValues = {
 };
 
 const std::vector<FakeQuantizeTransformationParam> fakeQuantizeOnDataValues = {
-        {
-                {256ul, {}, {0.f}, {2.55f}, {0.f}, {2.55f}},
-                "Pooling", "U8"
-        },
-        {
-                { 256ul, { 1ul }, { 0.f }, { 2.55f }, { 0.f }, { 2.55f } },
-                "Pooling", "U8"
-        },
-        {
-                { 256ul, {}, { 0.f }, { 2.55f }, { -1.28f }, { 1.27f } },
-                "Pooling", "I8"
-        },
-        {
-                { 256ul, {}, { 0.f }, { 2.55f }, { 2.55f }, { 2.55f } },
-                "Pooling", "U8"
-        },
-        {
-                { 16ul, {}, { 0.f }, { 1.5f }, { 0.f }, { 1.5f } },
-                "Pooling", "FP32"
-        },
-        {
-                { 16ul, {}, { -8.f }, { 7.f }, { -0.8f }, { 0.7f } },
-                "Pooling", "FP32"
-        },
+    {
+        {256ul, {}, {0.f}, {2.55f}, {0.f}, {2.55f}},
+        "Pooling", "U8"
+    },
+    {
+        { 256ul, { 1ul }, { 0.f }, { 2.55f }, { 0.f }, { 2.55f } },
+        "Pooling", "U8"
+    },
+    {
+        { 256ul, {}, { 0.f }, { 2.55f }, { -1.28f }, { 1.27f } },
+        "Pooling", "I8"
+    },
+    {
+        { 256ul, {}, { 0.f }, { 2.55f }, { 2.55f }, { 2.55f } },
+        "Pooling", "U8"
+    },
+    {
+        { 256ul, {}, { -127.5f }, { 0.f }, { -127.5f }, { 0.f } },
+        "Pooling", "U8"
+    },
+    {
+        { 16ul, {}, { 0.f }, { 1.5f }, { 0.f }, { 1.5f } },
+        "Pooling", "FP32"
+    },
+    {
+        { 16ul, {}, { -8.f }, { 7.f }, { -0.8f }, { 0.7f } },
+        "Pooling", "FP32"
+    },
     // nGraph: I8->FP32 Convert is not supported
     // { 256ul, {}, { -1.28f} , { 1.27f }, { -1.28f} , { 1.27f } },
     // { 256ul, { 1ul }, { -1.28f} , { 1.27f } }