diff --git a/source/application/main/ImageUtils.cc b/source/application/main/ImageUtils.cc
new file mode 100644
index 0000000..31b9493
--- /dev/null
+++ b/source/application/main/ImageUtils.cc
@@ -0,0 +1,126 @@
+/*
+ * Copyright (c) 2022 Arm Limited. All rights reserved.
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+#include "ImageUtils.hpp"
+
+#include <limits>
+
+namespace arm {
+namespace app {
+namespace image {
+
+    float Calculate1DOverlap(float x1Center, float width1, float x2Center, float width2)
+    {
+        float left_1 = x1Center - width1/2;
+        float left_2 = x2Center - width2/2;
+        float leftest = left_1 > left_2 ? left_1 : left_2;
+
+        float right_1 = x1Center + width1/2;
+        float right_2 = x2Center + width2/2;
+        float rightest = right_1 < right_2 ? right_1 : right_2;
+
+        return rightest - leftest;
+    }
+
+    float CalculateBoxIntersect(Box& box1, Box& box2)
+    {
+        float width = Calculate1DOverlap(box1.x, box1.w, box2.x, box2.w);
+        if (width < 0) {
+            return 0;
+        }
+        float height = Calculate1DOverlap(box1.y, box1.h, box2.y, box2.h);
+        if (height < 0) {
+            return 0;
+        }
+
+        float total_area = width*height;
+        return total_area;
+    }
+
+    float CalculateBoxUnion(Box& box1, Box& box2)
+    {
+        float boxes_intersection = CalculateBoxIntersect(box1, box2);
+        float boxes_union = box1.w * box1.h + box2.w * box2.h - boxes_intersection;
+        return boxes_union;
+    }
+
+    float CalculateBoxIOU(Box& box1, Box& box2)
+    {
+        float boxes_intersection = CalculateBoxIntersect(box1, box2);
+        if (boxes_intersection == 0) {
+            return 0;
+        }
+
+        float boxes_union = CalculateBoxUnion(box1, box2);
+        if (boxes_union == 0) {
+            return 0;
+        }
+
+        return boxes_intersection / boxes_union;
+    }
+
+    void CalculateNMS(std::forward_list<Detection>& detections, int classes, float iouThreshold)
+    {
+        int idxClass{0};
+        auto CompareProbs = [idxClass](Detection& prob1, Detection& prob2) {
+            return prob1.prob[idxClass] > prob2.prob[idxClass];
+        };
+
+        for (idxClass = 0; idxClass < classes; ++idxClass) {
+            detections.sort(CompareProbs);
+
+            for (auto it=detections.begin(); it != detections.end(); ++it) {
+                if (it->prob[idxClass] == 0) continue;
+                for (auto itc=std::next(it, 1); itc != detections.end(); ++itc) {
+                    if (itc->prob[idxClass] == 0) {
+                        continue;
+                    }
+                    if (CalculateBoxIOU(it->bbox, itc->bbox) > iouThreshold) {
+                        itc->prob[idxClass] = 0;
+                    }
+                }
+            }
+        }
+    }
+
+    void ConvertImgToInt8(void* data, const size_t kMaxImageSize)
+    {
+        auto* tmp_req_data = static_cast<uint8_t*>(data);
+        auto* tmp_signed_req_data = static_cast<int8_t*>(data);
+
+        for (size_t i = 0; i < kMaxImageSize; i++) {
+            tmp_signed_req_data[i] = (int8_t) (
+                    (int32_t) (tmp_req_data[i]) - 128);
+        }
+    }
+
+    void RgbToGrayscale(const uint8_t* srcPtr, uint8_t* dstPtr, const size_t dstImgSz)
+    {
+        const float R = 0.299;
+        const float G = 0.587;
+        const float B = 0.114;
+        for (size_t i = 0; i < dstImgSz; ++i, srcPtr += 3) {
+            uint32_t  int_gray = R * (*srcPtr) +
+                                 G * (*(srcPtr + 1)) +
+                                 B * (*(srcPtr + 2));
+            *dstPtr++ = int_gray <= std::numeric_limits<uint8_t>::max() ?
+                        int_gray : std::numeric_limits<uint8_t>::max();
+        }
+    }
+
+} /* namespace image */
+} /* namespace app */
+} /* namespace arm */
\ No newline at end of file
diff --git a/source/application/main/UseCaseCommonUtils.cc b/source/application/main/UseCaseCommonUtils.cc
index dd9a32d..d439446 100644
--- a/source/application/main/UseCaseCommonUtils.cc
+++ b/source/application/main/UseCaseCommonUtils.cc
@@ -15,6 +15,7 @@
  * limitations under the License.
  */
 #include "UseCaseCommonUtils.hpp"
+#include "ImageUtils.hpp"
 #include "InputFiles.hpp"
 #include "log_macros.h"
 
@@ -35,19 +36,10 @@
     fflush(stdout);
 }
 
-void image::ConvertImgToInt8(void* data, const size_t kMaxImageSize)
-{
-    auto* tmp_req_data = static_cast<uint8_t *>(data);
-    auto* tmp_signed_req_data = static_cast<int8_t *>(data);
-
-    for (size_t i = 0; i < kMaxImageSize; i++) {
-        tmp_signed_req_data[i] = (int8_t) (
-            (int32_t) (tmp_req_data[i]) - 128);
-    }
-}
 
 
-bool image::PresentInferenceResult(
+
+bool PresentInferenceResult(
     hal_platform &platform,
     const std::vector<arm::app::ClassificationResult> &results)
 {
@@ -93,19 +85,6 @@
     return true;
 }
 
-void image::RgbToGrayscale(const uint8_t *srcPtr, uint8_t *dstPtr, const size_t dstImgSz)
-{
-    float R=0.299;
-    float G=0.587;
-    float B=0.114;
-    for (size_t i = 0; i < dstImgSz; ++i, srcPtr += 3) {
-        uint32_t  int_gray = R * (*srcPtr) +
-                             G * (*(srcPtr + 1)) +
-                             B * (*(srcPtr + 2));
-        *dstPtr++ = int_gray <= std::numeric_limits<uint8_t>::max() ?
-                        int_gray : std::numeric_limits<uint8_t>::max();
-    }
-}
 
 void IncrementAppCtxIfmIdx(arm::app::ApplicationContext& ctx, const std::string& useCase)
 {
diff --git a/source/application/main/include/ImageUtils.hpp b/source/application/main/include/ImageUtils.hpp
new file mode 100644
index 0000000..a8c7650
--- /dev/null
+++ b/source/application/main/include/ImageUtils.hpp
@@ -0,0 +1,116 @@
+/*
+ * Copyright (c) 2022 Arm Limited. All rights reserved.
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+#ifndef IMAGE_UTILS_HPP
+#define IMAGE_UTILS_HPP
+
+#include <cstddef>
+#include <cstdint>
+#include <forward_list>
+#include <vector>
+
+/* Helper macro to convert RGB888 to RGB565 format. */
+#define RGB888_TO_RGB565(R8,G8,B8)  ((((R8>>3) & 0x1F) << 11) |     \
+                                     (((G8>>2) & 0x3F) << 5)  |     \
+                                     ((B8>>3) & 0x1F))
+
+constexpr uint16_t COLOR_BLACK  = 0;
+constexpr uint16_t COLOR_GREEN  = RGB888_TO_RGB565(  0, 255,  0); // 2016;
+constexpr uint16_t COLOR_YELLOW = RGB888_TO_RGB565(255, 255,  0); // 65504;
+
+
+namespace arm {
+namespace app {
+namespace image {
+
+    /**
+     * Contains the x,y co-ordinates of a box centre along with the box width and height.
+     */
+    struct Box {
+        float x;
+        float y;
+        float w;
+        float h;
+    };
+
+    struct Detection {
+        Box bbox;
+        std::vector<float> prob;
+        float objectness;
+    };
+
+    /**
+     * @brief       Calculate the 1D overlap.
+     * @param[in]   x1Center   First center point.
+     * @param[in]   width1     First width.
+     * @param[in]   x2Center   Second center point.
+     * @param[in]   width2     Second width.
+     * @return      The overlap between the two lines.
+     **/
+    float Calculate1DOverlap(float x1Center, float width1, float x2Center, float width2);
+
+    /**
+     * @brief       Calculate the intersection between the two given boxes.
+     * @param[in]   box1   First box.
+     * @param[in]   box2   Second box.
+     * @return      The intersection value.
+     **/
+    float CalculateBoxIntersect(Box& box1, Box& box2);
+
+    /**
+     * @brief       Calculate the union between the two given boxes.
+     * @param[in]   box1   First box.
+     * @param[in]   box2   Second box.
+     * @return      The two given boxes union value.
+     **/
+    float CalculateBoxUnion(Box& box1, Box& box2);
+
+    /**
+     * @brief       Calculate the intersection over union between the two given boxes.
+     * @param[in]   box1   First box.
+     * @param[in]   box2   Second box.
+     * @return      The intersection over union value.
+     **/
+    float CalculateBoxIOU(Box& box1, Box& box2);
+
+    /**
+     * @brief       Calculate the Non-Maxima suppression on the given detection boxes.
+     * @param[in]   detections    List of Detection boxes.
+     * @param[in]   classes       Number of classes.
+     * @param[in]   iouThreshold  Intersection over union threshold.
+     **/
+    void CalculateNMS(std::forward_list<Detection>& detections, int classes, float iouThreshold);
+
+    /**
+     * @brief           Helper function to convert a UINT8 image to INT8 format.
+     * @param[in,out]   data            Pointer to the data start.
+     * @param[in]       kMaxImageSize   Total number of pixels in the image.
+     **/
+    void ConvertImgToInt8(void* data, size_t kMaxImageSize);
+
+    /**
+     * @brief       Converts RGB image to grayscale.
+     * @param[in]   srcPtr   Pointer to RGB source image.
+     * @param[out]  dstPtr   Pointer to grayscale destination image.
+     * @param[in]   imgSz    Destination image size.
+     **/
+    void RgbToGrayscale(const uint8_t* srcPtr, uint8_t* dstPtr, size_t dstImgSz);
+
+} /* namespace image */
+} /* namespace app */
+} /* namespace arm */
+
+#endif /* IMAGE_UTILS_HPP */
\ No newline at end of file
diff --git a/source/application/main/include/UseCaseCommonUtils.hpp b/source/application/main/include/UseCaseCommonUtils.hpp
index cd0cb69..7f5dde6 100644
--- a/source/application/main/include/UseCaseCommonUtils.hpp
+++ b/source/application/main/include/UseCaseCommonUtils.hpp
@@ -24,30 +24,10 @@
 #include "UseCaseHandler.hpp"       /* Handlers for different user options. */
 #include "Classifier.hpp"           /* Classifier. */
 #include "InputFiles.hpp"
-#include <cinttypes>
-
-
-/* Helper macro to convert RGB888 to RGB565 format. */
-#define RGB888_TO_RGB565(R8,G8,B8)  ((((R8>>3) & 0x1F) << 11) |     \
-                                     (((G8>>2) & 0x3F) << 5)  |     \
-                                     ((B8>>3) & 0x1F))
-
-constexpr uint16_t COLOR_BLACK  = 0;
-constexpr uint16_t COLOR_GREEN  = RGB888_TO_RGB565(  0u, 255u,  0u); // 2016;
-constexpr uint16_t COLOR_YELLOW = RGB888_TO_RGB565(255u, 255u,  0u); // 65504;
 
 
 void DisplayCommonMenu();
 
-namespace image{
-
-  /**
-  * @brief           Helper function to convert a UINT8 image to INT8 format.
-  * @param[in,out]   data            Pointer to the data start.
-  * @param[in]       kMaxImageSize   Total number of pixels in the image.
-  **/
-  void ConvertImgToInt8(void * data, size_t kMaxImageSize);
-
   /**
    * @brief           Presents inference results using the data presentation
    *                  object.
@@ -55,17 +35,9 @@
    * @param[in]       results     Vector of classification results to be displayed.
    * @return          true if successful, false otherwise.
    **/
-  bool PresentInferenceResult(hal_platform & platform,
-    const std::vector < arm::app::ClassificationResult > & results);
+bool PresentInferenceResult(hal_platform& platform,
+                            const std::vector<arm::app::ClassificationResult>& results);
 
-  /**
-   * @brief       Converts RGB image to grayscale.
-   * @param[in]   srcPtr   Pointer to RGB source image.
-   * @param[out]  dstPtr   Pointer to grayscale destination image.
-   * @param[in]   imgSz    Destination image size.
-   **/
-  void RgbToGrayscale(const uint8_t *srcPtr, uint8_t *dstPtr, const size_t dstImgSz);
-}
 
 /**
    * @brief           Helper function to increment current input feature vector index.
diff --git a/source/math/PlatformMath.cc b/source/math/PlatformMath.cc
index cc603f3..b666d29 100644
--- a/source/math/PlatformMath.cc
+++ b/source/math/PlatformMath.cc
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2021 Arm Limited. All rights reserved.
+ * Copyright (c) 2021-2022 Arm Limited. All rights reserved.
  * SPDX-License-Identifier: Apache-2.0
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
@@ -303,6 +303,11 @@
         }
     }
 
+    float MathUtils::SigmoidF32(float x)
+    {
+        return 1.f/(1.f + std::exp(-x));
+    }
+
 } /* namespace math */
 } /* namespace app */
 } /* namespace arm */
diff --git a/source/math/include/PlatformMath.hpp b/source/math/include/PlatformMath.hpp
index 5ac10de..2bf7733 100644
--- a/source/math/include/PlatformMath.hpp
+++ b/source/math/include/PlatformMath.hpp
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2021 Arm Limited. All rights reserved.
+ * Copyright (c) 2021-2022 Arm Limited. All rights reserved.
  * SPDX-License-Identifier: Apache-2.0
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
@@ -19,7 +19,6 @@
 
 /* See if ARM DSP functions can be used. */
 #if defined(ARM_MATH_DSP)
-
     #include "arm_math.h"
     #define M_PI    (PI)
 #else
@@ -127,8 +126,8 @@
          * @param[out]  output   Pre-allocated buffer to be populated with
          *                       natural log values of each input element.
          */
-        static void VecLogarithmF32(std::vector <float>& input,
-                                    std::vector <float>& output);
+        static void VecLogarithmF32(std::vector<float>& input,
+                                    std::vector<float>& output);
 
         /**
          * @brief       Computes the dot product of two 1D floating point
@@ -165,6 +164,13 @@
         * @param[in]   vector Vector of floats modified in-place
         */
         static void SoftmaxF32(std::vector<float>& vec);
+
+        /**
+        * @brief       Calculate the Sigmoid function of the given value.
+        * @param[in]   x   Value to apply Sigmoid to.
+        * @return      Sigmoid value of the input.
+        */
+        static float SigmoidF32(float x);
     };
 
 } /* namespace math */
diff --git a/source/use_case/ad/src/UseCaseHandler.cc b/source/use_case/ad/src/UseCaseHandler.cc
index 420e6d4..853ab08 100644
--- a/source/use_case/ad/src/UseCaseHandler.cc
+++ b/source/use_case/ad/src/UseCaseHandler.cc
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2021 Arm Limited. All rights reserved.
+ * Copyright (c) 2021-2022 Arm Limited. All rights reserved.
  * SPDX-License-Identifier: Apache-2.0
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
@@ -22,6 +22,7 @@
 #include "hal.h"
 #include "AdMelSpectrogram.hpp"
 #include "AudioUtils.hpp"
+#include "ImageUtils.hpp"
 #include "UseCaseCommonUtils.hpp"
 #include "AdPostProcessing.hpp"
 #include "log_macros.h"
diff --git a/source/use_case/asr/src/UseCaseHandler.cc b/source/use_case/asr/src/UseCaseHandler.cc
index afcb6e4..7bce2c6 100644
--- a/source/use_case/asr/src/UseCaseHandler.cc
+++ b/source/use_case/asr/src/UseCaseHandler.cc
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2021 Arm Limited. All rights reserved.
+ * Copyright (c) 2021-2022 Arm Limited. All rights reserved.
  * SPDX-License-Identifier: Apache-2.0
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
@@ -22,6 +22,7 @@
 #include "hal.h"
 #include "Wav2LetterMfcc.hpp"
 #include "AudioUtils.hpp"
+#include "ImageUtils.hpp"
 #include "UseCaseCommonUtils.hpp"
 #include "AsrResult.hpp"
 #include "Wav2LetterPreprocess.hpp"
diff --git a/source/use_case/img_class/src/UseCaseHandler.cc b/source/use_case/img_class/src/UseCaseHandler.cc
index fafc6b9..1f1d78b 100644
--- a/source/use_case/img_class/src/UseCaseHandler.cc
+++ b/source/use_case/img_class/src/UseCaseHandler.cc
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2021 Arm Limited. All rights reserved.
+ * Copyright (c) 2021-2022 Arm Limited. All rights reserved.
  * SPDX-License-Identifier: Apache-2.0
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
@@ -19,6 +19,7 @@
 #include "Classifier.hpp"
 #include "InputFiles.hpp"
 #include "MobileNetModel.hpp"
+#include "ImageUtils.hpp"
 #include "UseCaseCommonUtils.hpp"
 #include "hal.h"
 #include "log_macros.h"
@@ -136,7 +137,7 @@
             arm::app::DumpTensor(outputTensor);
 #endif /* VERIFY_TEST_OUTPUT */
 
-            if (!image::PresentInferenceResult(platform, results)) {
+            if (!PresentInferenceResult(platform, results)) {
                 return false;
             }
 
diff --git a/source/use_case/kws/src/UseCaseHandler.cc b/source/use_case/kws/src/UseCaseHandler.cc
index c2d2ea4..8dd7724 100644
--- a/source/use_case/kws/src/UseCaseHandler.cc
+++ b/source/use_case/kws/src/UseCaseHandler.cc
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2021 Arm Limited. All rights reserved.
+ * Copyright (c) 2021-2022 Arm Limited. All rights reserved.
  * SPDX-License-Identifier: Apache-2.0
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
@@ -22,6 +22,7 @@
 #include "hal.h"
 #include "MicroNetKwsMfcc.hpp"
 #include "AudioUtils.hpp"
+#include "ImageUtils.hpp"
 #include "UseCaseCommonUtils.hpp"
 #include "KwsResult.hpp"
 #include "log_macros.h"
diff --git a/source/use_case/kws_asr/src/UseCaseHandler.cc b/source/use_case/kws_asr/src/UseCaseHandler.cc
index bfc1d25..d598de6 100644
--- a/source/use_case/kws_asr/src/UseCaseHandler.cc
+++ b/source/use_case/kws_asr/src/UseCaseHandler.cc
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2021 Arm Limited. All rights reserved.
+ * Copyright (c) 2021-2022 Arm Limited. All rights reserved.
  * SPDX-License-Identifier: Apache-2.0
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
@@ -19,6 +19,7 @@
 #include "hal.h"
 #include "InputFiles.hpp"
 #include "AudioUtils.hpp"
+#include "ImageUtils.hpp"
 #include "UseCaseCommonUtils.hpp"
 #include "MicroNetKwsModel.hpp"
 #include "MicroNetKwsMfcc.hpp"
diff --git a/source/use_case/noise_reduction/src/UseCaseHandler.cc b/source/use_case/noise_reduction/src/UseCaseHandler.cc
index 0c5984c..792b460 100644
--- a/source/use_case/noise_reduction/src/UseCaseHandler.cc
+++ b/source/use_case/noise_reduction/src/UseCaseHandler.cc
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2021 Arm Limited. All rights reserved.
+ * Copyright (c) 2021-2022 Arm Limited. All rights reserved.
  * SPDX-License-Identifier: Apache-2.0
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
@@ -18,6 +18,7 @@
 #include "UseCaseHandler.hpp"
 #include "UseCaseCommonUtils.hpp"
 #include "AudioUtils.hpp"
+#include "ImageUtils.hpp"
 #include "InputFiles.hpp"
 #include "RNNoiseModel.hpp"
 #include "RNNoiseProcess.hpp"
diff --git a/source/use_case/object_detection/include/DetectorPostProcessing.hpp b/source/use_case/object_detection/include/DetectorPostProcessing.hpp
index 3e9c819..5393f89 100644
--- a/source/use_case/object_detection/include/DetectorPostProcessing.hpp
+++ b/source/use_case/object_detection/include/DetectorPostProcessing.hpp
@@ -18,6 +18,7 @@
 #define DETECTOR_POST_PROCESSING_HPP
 
 #include "UseCaseCommonUtils.hpp"
+#include "ImageUtils.hpp"
 #include "DetectionResult.hpp"
 #include "YoloFastestModel.hpp"
 
@@ -45,20 +46,6 @@
         int topN;
     };
 
-
-    struct Box {
-        float x;
-        float y;
-        float w;
-        float h;
-    };
-
-    struct Detection {
-        Box bbox;
-        std::vector<float> prob;
-        float objectness;
-    };
-
     /**
      * @brief   Helper class to manage tensor post-processing for "object_detection"
      *          output.
@@ -99,18 +86,11 @@
         int   m_topN;       /* TopN */
 
         /**
-         * @brief       Calculate the Sigmoid function of the give value.
-         * @param[in]   x   Value.
-         * @return      Sigmoid value of the input.
-         **/
-        float Sigmoid(float x);
-
-        /**
          * @brief       Insert the given Detection in the list.
          * @param[in]   detections   List of detections.
          * @param[in]   det          Detection to be inserted.
          **/
-        void InsertTopNDetections(std::forward_list<Detection>& detections, Detection& det);
+        void InsertTopNDetections(std::forward_list<image::Detection>& detections, image::Detection& det);
 
         /**
          * @brief        Given a Network calculate the detection boxes.
@@ -124,49 +104,7 @@
                              int imageWidth,
                              int imageHeight,
                              float threshold,
-                             std::forward_list<Detection>& detections);
-
-        /**
-         * @brief       Calculate the 1D overlap.
-         * @param[in]   x1Center   First center point.
-         * @param[in]   width1     First width.
-         * @param[in]   x2Center   Second center point.
-         * @param[in]   width2     Second width.
-         * @return      The overlap between the two lines.
-         **/
-        float Calculate1DOverlap(float x1Center, float width1, float x2Center, float width2);
-
-        /**
-         * @brief       Calculate the intersection between the two given boxes.
-         * @param[in]   box1   First box.
-         * @param[in]   box2   Second box.
-         * @return      The intersection value.
-         **/
-        float CalculateBoxIntersect(Box& box1, Box& box2);
-
-        /**
-         * @brief       Calculate the union between the two given boxes.
-         * @param[in]   box1   First box.
-         * @param[in]   box2   Second box.
-         * @return      The two given boxes union value.
-         **/
-        float CalculateBoxUnion(Box& box1, Box& box2);
-        /**
-         * @brief       Calculate the intersection over union between the two given boxes.
-         * @param[in]   box1   First box.
-         * @param[in]   box2   Second box.
-         * @return      The intersection over union value.
-         **/
-        float CalculateBoxIOU(Box& box1, Box& box2);
-
-        /**
-         * @brief       Calculate the Non-Maxima suppression on the given detection boxes.
-         * @param[in]   detections    Detection boxes.
-         * @param[in]   classes       Number of classes.
-         * @param[in]   iouThreshold  Intersection over union threshold.
-         * @return      true or false based on execution success.
-         **/
-        void  CalculateNMS(std::forward_list<Detection>& detections, int classes, float iouThreshold);
+                             std::forward_list<image::Detection>& detections);
 
         /**
          * @brief       Draw on the given image a bounding box starting at (boxX, boxY).
diff --git a/source/use_case/object_detection/src/DetectorPostProcessing.cc b/source/use_case/object_detection/src/DetectorPostProcessing.cc
index edfb137..e97e6b3 100644
--- a/source/use_case/object_detection/src/DetectorPostProcessing.cc
+++ b/source/use_case/object_detection/src/DetectorPostProcessing.cc
@@ -15,6 +15,7 @@
  * limitations under the License.
  */
 #include "DetectorPostProcessing.hpp"
+#include "PlatformMath.hpp"
 
 #include <algorithm>
 #include <cmath>
@@ -75,7 +76,7 @@
     int originalImageWidth = originalImageSize;
     int originalImageHeight = originalImageSize;
 
-    std::forward_list<Detection> detections;
+    std::forward_list<image::Detection> detections;
     GetNetworkBoxes(net, originalImageWidth, originalImageHeight, m_threshold, detections);
 
     /* Do nms */
@@ -124,15 +125,11 @@
     }
 }
 
-float DetectorPostprocessing::Sigmoid(float x)
-{
-    return 1.f/(1.f + exp(-x));
-}
 
-void DetectorPostprocessing::InsertTopNDetections(std::forward_list<Detection>& detections, Detection& det)
+void DetectorPostprocessing::InsertTopNDetections(std::forward_list<image::Detection>& detections, image::Detection& det)
 {
-    std::forward_list<Detection>::iterator it;
-    std::forward_list<Detection>::iterator last_it;
+    std::forward_list<image::Detection>::iterator it;
+    std::forward_list<image::Detection>::iterator last_it;
     for ( it = detections.begin(); it != detections.end(); ++it ) {
         if(it->objectness > det.objectness)
             break;
@@ -144,11 +141,11 @@
     }
 }
 
-void DetectorPostprocessing::GetNetworkBoxes(Network& net, int imageWidth, int imageHeight, float threshold, std::forward_list<Detection>& detections)
+void DetectorPostprocessing::GetNetworkBoxes(Network& net, int imageWidth, int imageHeight, float threshold, std::forward_list<image::Detection>& detections)
 {
     int numClasses = net.numClasses;
     int num = 0;
-    auto det_objectness_comparator = [](Detection& pa, Detection& pb) {
+    auto det_objectness_comparator = [](image::Detection& pa, image::Detection& pb) {
         return pa.objectness < pb.objectness;
     };
     for (size_t i = 0; i < net.branches.size(); ++i) {
@@ -162,10 +159,10 @@
 
                     /* Objectness score */
                     int bbox_obj_offset = h * width * channel + w * channel + anc * (numClasses + 5) + 4;
-                    float objectness = Sigmoid(((float)net.branches[i].modelOutput[bbox_obj_offset] - net.branches[i].zeroPoint) * net.branches[i].scale);
+                    float objectness = math::MathUtils::SigmoidF32(((float)net.branches[i].modelOutput[bbox_obj_offset] - net.branches[i].zeroPoint) * net.branches[i].scale);
 
                     if(objectness > threshold) {
-                        Detection det;
+                        image::Detection det;
                         det.objectness = objectness;
                         /* Get bbox prediction data for each anchor, each feature point */
                         int bbox_x_offset = bbox_obj_offset -4;
@@ -183,8 +180,8 @@
                         float bbox_x, bbox_y;
 
                         /* Eliminate grid sensitivity trick involved in YOLOv4 */
-                        bbox_x = Sigmoid(det.bbox.x);
-                        bbox_y = Sigmoid(det.bbox.y);
+                        bbox_x = math::MathUtils::SigmoidF32(det.bbox.x);
+                        bbox_y = math::MathUtils::SigmoidF32(det.bbox.y);
                         det.bbox.x = (bbox_x + w) / width;
                         det.bbox.y = (bbox_y + h) / height;
 
@@ -192,7 +189,7 @@
                         det.bbox.h = exp(det.bbox.h) * net.branches[i].anchor[anc*2+1] / net.inputHeight;
 
                         for (int s = 0; s < numClasses; s++) {
-                            float sig = Sigmoid(((float)net.branches[i].modelOutput[bbox_scores_offset + s] - net.branches[i].zeroPoint) * net.branches[i].scale)*objectness;
+                            float sig = math::MathUtils::SigmoidF32(((float)net.branches[i].modelOutput[bbox_scores_offset + s] - net.branches[i].zeroPoint) * net.branches[i].scale)*objectness;
                             det.prob.emplace_back((sig > threshold) ? sig : 0);
                         }
 
@@ -221,81 +218,6 @@
         num -=1;
 }
 
-float DetectorPostprocessing::Calculate1DOverlap(float x1Center, float width1, float x2Center, float width2)
-{
-    float left_1 = x1Center - width1/2;
-    float left_2 = x2Center - width2/2;
-    float leftest = left_1 > left_2 ? left_1 : left_2;
-
-    float right_1 = x1Center + width1/2;
-    float right_2 = x2Center + width2/2;
-    float rightest = right_1 < right_2 ? right_1 : right_2;
-
-    return rightest - leftest;
-}
-
-float DetectorPostprocessing::CalculateBoxIntersect(Box& box1, Box& box2)
-{
-    float width = Calculate1DOverlap(box1.x, box1.w, box2.x, box2.w);
-    if (width < 0) {
-        return 0;
-    }
-    float height = Calculate1DOverlap(box1.y, box1.h, box2.y, box2.h);
-    if (height < 0) {
-        return 0;
-    }
-
-    float total_area = width*height;
-    return total_area;
-}
-
-float DetectorPostprocessing::CalculateBoxUnion(Box& box1, Box& box2)
-{
-    float boxes_intersection = CalculateBoxIntersect(box1, box2);
-    float boxes_union = box1.w * box1.h + box2.w * box2.h - boxes_intersection;
-    return boxes_union;
-}
-
-
-float DetectorPostprocessing::CalculateBoxIOU(Box& box1, Box& box2)
-{
-    float boxes_intersection = CalculateBoxIntersect(box1, box2);
-    if (boxes_intersection == 0) {
-        return 0;
-    }
-
-    float boxes_union = CalculateBoxUnion(box1, box2);
-    if (boxes_union == 0) {
-        return 0;
-    }
-
-    return boxes_intersection / boxes_union;
-}
-
-void DetectorPostprocessing::CalculateNMS(std::forward_list<Detection>& detections, int classes, float iouThreshold)
-{
-    int idxClass{0};
-    auto CompareProbs = [idxClass](Detection& prob1, Detection& prob2) {
-        return prob1.prob[idxClass] > prob2.prob[idxClass];
-    };
-
-    for (idxClass = 0; idxClass < classes; ++idxClass) {
-        detections.sort(CompareProbs);
-
-        for (std::forward_list<Detection>::iterator it=detections.begin(); it != detections.end(); ++it) {
-            if (it->prob[idxClass] == 0) continue;
-            for (std::forward_list<Detection>::iterator itc=std::next(it, 1); itc != detections.end(); ++itc) {
-                if (itc->prob[idxClass] == 0) {
-                    continue;
-                }
-                if (CalculateBoxIOU(it->bbox, itc->bbox) > iouThreshold) {
-                    itc->prob[idxClass] = 0;
-                }
-            }
-        }
-    }
-}
-
 void DetectorPostprocessing::DrawBoxOnImage(uint8_t* imgIn, int imWidth, int imHeight, int boxX,int boxY, int boxWidth, int boxHeight)
 {
     auto CheckAndFixOffset = [](int im_width,int im_height,int& offset) {
diff --git a/source/use_case/vww/src/UseCaseHandler.cc b/source/use_case/vww/src/UseCaseHandler.cc
index 01011e2..a47f191 100644
--- a/source/use_case/vww/src/UseCaseHandler.cc
+++ b/source/use_case/vww/src/UseCaseHandler.cc
@@ -18,6 +18,7 @@
 #include "VisualWakeWordModel.hpp"
 #include "Classifier.hpp"
 #include "InputFiles.hpp"
+#include "ImageUtils.hpp"
 #include "UseCaseCommonUtils.hpp"
 #include "hal.h"
 #include "log_macros.h"
@@ -144,7 +145,7 @@
             arm::app::DumpTensor(outputTensor);
 #endif /* VERIFY_TEST_OUTPUT */
 
-            if (!image::PresentInferenceResult(platform, results)) {
+            if (!PresentInferenceResult(platform, results)) {
                 return false;
             }
 
diff --git a/tests/use_case/img_class/InferenceTestMobilenetV2.cc b/tests/use_case/img_class/InferenceTestMobilenetV2.cc
index 294215f..7e7508b 100644
--- a/tests/use_case/img_class/InferenceTestMobilenetV2.cc
+++ b/tests/use_case/img_class/InferenceTestMobilenetV2.cc
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2021 Arm Limited. All rights reserved.
+ * Copyright (c) 2021-2022 Arm Limited. All rights reserved.
  * SPDX-License-Identifier: Apache-2.0
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
@@ -34,7 +34,7 @@
     memcpy(inputTensor->data.data, imageData, copySz);
 
     if(model.IsDataSigned()){
-        convertImgIoInt8(inputTensor->data.data, copySz);
+        arm::app::image::ConvertImgToInt8(inputTensor->data.data, copySz);
     }
 
     return model.RunInference();
diff --git a/tests/use_case/object_detection/InferenceTestYoloFastest.cc b/tests/use_case/object_detection/InferenceTestYoloFastest.cc
index d9cabbd..b3bd408 100644
--- a/tests/use_case/object_detection/InferenceTestYoloFastest.cc
+++ b/tests/use_case/object_detection/InferenceTestYoloFastest.cc
@@ -66,10 +66,10 @@
     const size_t copySz = inputTensor->bytes < IMAGE_DATA_SIZE ?
                             inputTensor->bytes : IMAGE_DATA_SIZE;
 
-    image::RgbToGrayscale(imageData,inputTensor->data.uint8,copySz);
+    arm::app::image::RgbToGrayscale(imageData,inputTensor->data.uint8,copySz);
 
     if(model.IsDataSigned()){
-        convertImgIoInt8(inputTensor->data.data, copySz);
+        arm::app::image::ConvertImgToInt8(inputTensor->data.data, copySz);
     }
 
     return model.RunInference();
diff --git a/tests/use_case/vww/InferenceVisualWakeWordModelTests.cc b/tests/use_case/vww/InferenceVisualWakeWordModelTests.cc
index 04dce0d..194099f 100644
--- a/tests/use_case/vww/InferenceVisualWakeWordModelTests.cc
+++ b/tests/use_case/vww/InferenceVisualWakeWordModelTests.cc
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2021 Arm Limited. All rights reserved.
+ * Copyright (c) 2021-2022 Arm Limited. All rights reserved.
  * SPDX-License-Identifier: Apache-2.0
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
@@ -34,7 +34,7 @@
     memcpy(inputTensor->data.data, imageData, copySz);
 
     if(model.IsDataSigned()){
-        convertImgIoInt8(inputTensor->data.data, copySz);
+        arm::app::image::ConvertImgToInt8(inputTensor->data.data, copySz);
     }
 
     return model.RunInference();
diff --git a/tests/utils/ImageUtils.cc b/tests/utils/ImageUtils.cc
deleted file mode 100644
index 506040f..0000000
--- a/tests/utils/ImageUtils.cc
+++ /dev/null
@@ -1,37 +0,0 @@
-/*
- * Copyright (c) 2021 Arm Limited. All rights reserved.
- * SPDX-License-Identifier: Apache-2.0
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *     http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- */
-#include "ImageUtils.hpp"
-
-void convertImgIoInt8(void * data, const size_t sz)
-{
-    uint8_t * tmp_req_data          = static_cast<uint8_t *>(data);
-    int8_t * tmp_signed_req_data    = static_cast<int8_t *>(data);
-
-    for (size_t i = 0; i < sz; ++i) {
-        tmp_signed_req_data[i] = static_cast<int8_t>(
-                static_cast<int32_t>(tmp_req_data[i]) - 128);
-    }
-}
-
-void convertImgIoGreyscale(const uint8_t * srcPtr, uint8_t * dstPtr, const size_t sz)
-{
-    for (size_t i = 0; i < sz; ++i, srcPtr += 3) {
-        *dstPtr++ = 0.2989 * (*srcPtr) +
-                    0.587 * (*(srcPtr+1)) +
-                    0.114 * (*(srcPtr+2));
-    }
-}
\ No newline at end of file
diff --git a/tests/utils/ImageUtils.hpp b/tests/utils/ImageUtils.hpp
deleted file mode 100644
index 838dcef..0000000
--- a/tests/utils/ImageUtils.hpp
+++ /dev/null
@@ -1,26 +0,0 @@
-/*
- * Copyright (c) 2021 Arm Limited. All rights reserved.
- * SPDX-License-Identifier: Apache-2.0
- *
- * Licensed under the Apache License, Version 2.0 (the "License");
- * you may not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- *     http://www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an "AS IS" BASIS,
- * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- */
-#ifndef IMAGEUTILS_HPP
-#define IMAGEUTILS_HPP
-
-#include <catch.hpp>
-
-void convertImgIoInt8(void * data, const size_t sz);
-
-void convertImgIoGreyscale(const uint8_t * srcPtr, uint8_t * dstPtr, const size_t sz);
-
-#endif /* IMAGEUTILS_HPP */