diff --git a/CHANGELOG b/CHANGELOG
index a39a220bd..dee448522 100644
--- a/CHANGELOG
+++ b/CHANGELOG
@@ -11,6 +11,7 @@ improved doc.
     - [spatialPartitioning] Add kd-tree traits type (#80)
     - [fitting] Add Primitive::getNumNeighbors (#86)
     - [fitting] Change naming convention and rational for principal curvatures in CurvatureEstimatorBase (#94)
+    - [fitting] Add class MeanNormalDer to compute derivatives of the mean normal vector (#78)
 
 - Bug-fixes and code improvements
     - [spatialPartitioning] Fix unwanted function hiding with DryFit::setWeightFunc (#86)
diff --git a/Ponca/src/Fitting/mean.h b/Ponca/src/Fitting/mean.h
index f037dc1ea..c70ab81cf 100644
--- a/Ponca/src/Fitting/mean.h
+++ b/Ponca/src/Fitting/mean.h
@@ -122,6 +122,56 @@ namespace Ponca {
 
     }; //class MeanPositionDer
 
+    template<class DataPoint, class _WFunctor, int DiffType, typename T>
+    class MeanNormalDer : public T {
+        PONCA_FITTING_DECLARE_DEFAULT_TYPES
+        PONCA_FITTING_DECLARE_DEFAULT_DER_TYPES
+
+    protected:
+        enum {
+            Check = Base::PROVIDES_PRIMITIVE_DERIVATIVE && 
+                    Base::PROVIDES_MEAN_NORMAL,
+            PROVIDES_MEAN_NORMAL_DERIVATIVE,  /*!< \brief Provides derivative of the mean normal*/
+        };
+
+        private:
+        /*! \brief Derivatives of the input normals of the input points vectors*/
+        VectorArray m_dSumN {VectorArray::Zero()};
+
+    public:
+
+        PONCA_EXPLICIT_CAST_OPERATORS_DER(MeanNormalDer,meanNormalDer)
+        PONCA_FITTING_DECLARE_INIT
+        PONCA_FITTING_DECLARE_ADDNEIGHBOR_DER
+
+    /// \brief Compute the derivative of the mean normal vector of the input points. 
+    /// 
+    /// ### Step-by-step derivation of the mean normal : 
+    ///  Given the definition of the mean normal \f$ n(\mathbf{x}) = \frac{\sum_i w_\mathbf{x}(\mathbf{n_i}) \mathbf{n_i}}{\sum_i w_\mathbf{x}(\mathbf{n_i})} \f$, 
+    /// where \f$\left[\mathbf{n_i} \in \text{neighborhood}(\mathbf{x})\right]\f$ are all the normal samples in \f$\mathbf{x}\f$'s neighborhood.
+    ///
+    /// We denote \f$ t(\mathbf{x}) = \sum_i w_\mathbf{x}(\mathbf{n_i}) \mathbf{n_i} \f$ and \f$ s(\mathbf{x}) = \sum_i w_\mathbf{x}(\mathbf{n_i})\f$, 
+    /// such that \f$ n(\mathbf{x}) = \frac{t(\mathbf{x})}{s(\mathbf{x})}\f$.
+    ///
+    /// By definition, \f$ n'(\mathbf{x}) = \frac{s(\mathbf{x})t'(\mathbf{x}) - t(\mathbf{x})s'(\mathbf{x})}{s(\mathbf{x})^2}\f$.
+    ///
+    /// Assuming the weight of each normal is dependent on the position, we have \f$ s'(\mathbf{x}) = \sum_i w'_\mathbf{x}(\mathbf{n_i}) \f$.
+    ///
+    /// We rewrite \f$ t(\mathbf{x}) = \sum u(\mathbf{x})v(\mathbf{x}) \f$, with \f$ u(\mathbf{x}) = w_\mathbf{x}(\mathbf{n_i}) \f$ and \f$ v(\mathbf{x}) = \mathbf{n_i} \f$.
+    ///
+    /// Assuming the normal vectors themselves do not change with position, \f$v(\mathbf{x})\f$ is constant, its derivative is null, 
+    /// and so \f$ t'(\mathbf{x}) = \sum_i u'(\mathbf{x}) v(\mathbf{x}) = \sum_i w'_\mathbf{x}(\mathbf{n_i}) \mathbf{n_i} \f$.
+    ///
+    /// Which leads to \f$ n'(\mathbf{x}) = \frac{\sum_i w'\mathbf{x}(\mathbf{n_i}) \mathbf{n_i} - n(\mathbf{x})\sum w'(\mathbf{x})}{\sum_i w\mathbf{x}(\mathbf{n_i})} \f$.
+    /// \note This code is not directly tested. 
+
+    PONCA_MULTIARCH inline const VectorArray& dMeanNormal() const
+    { 
+        return m_dSumN; 
+    }
+
+    }; //class MeanNormalDer
+
 #include "mean.hpp"
 
 } //namespace Ponca
diff --git a/Ponca/src/Fitting/mean.hpp b/Ponca/src/Fitting/mean.hpp
index 3369581b5..2f4e19ed9 100644
--- a/Ponca/src/Fitting/mean.hpp
+++ b/Ponca/src/Fitting/mean.hpp
@@ -68,3 +68,25 @@ MeanPositionDer<DataPoint, _WFunctor, DiffType, T>::addLocalNeighbor(Scalar w,
 
     return false;
 }
+
+
+template<class DataPoint, class _WFunctor, int DiffType, typename T>
+void
+MeanNormalDer<DataPoint, _WFunctor, DiffType, T>::init(const VectorType &_evalPos) {
+    Base::init(_evalPos);
+    m_dSumN.setZero();
+}
+
+template<class DataPoint, class _WFunctor, int DiffType, typename T>
+bool
+MeanNormalDer<DataPoint, _WFunctor, DiffType, T>::addLocalNeighbor(Scalar w,
+                                                                 const VectorType &localQ,
+                                                                 const DataPoint &attributes,
+                                                                 ScalarArray &dw) {
+    if (Base::addLocalNeighbor(w, localQ, attributes, dw)) {
+        m_dSumN += attributes.normal() * dw;
+        return true;
+    }
+
+    return false;
+}
diff --git a/doc/src/ponca.mdoc b/doc/src/ponca.mdoc
index faff972e6..40750b357 100644
--- a/doc/src/ponca.mdoc
+++ b/doc/src/ponca.mdoc
@@ -46,14 +46,14 @@
   \section ponca_credits_sec Credits
 
   \subsection ponca_credits_subsection_crew Developers and contributors
+  Main contributors
   - <b> Nicolas Mellado </b>: conception, implementation, documentation and examples <br/>
-  - <b> Aniket Agarwalla </b> (Google Summer of Code 2021) : add new fitting primitives <br/>
-  - <b> Matthieu Gomiero, Dorian Ferreira, Sacha Vincent </b>: port KdTree code in Spatial Partitioning module <br/>
-  - <b> Nicolas Dupont, Carl Delrieu, Dorian Verge</b>: add Wendland and Singular weight kernels, as described in \cite Alexa:2009:Hermite<br/>
   - <b> Thibault Lejemble </b>: implementation of several fitting primitives, differential estimators, and original
   version of the Spatial Partitioning code<br/>
-  - <b> Jules Vidal, Antoine Lafon </b>: bug fixes<br/>
-  - <b> Amael Marquez </b>: KdTree API improvements <br/>
+
+  Other contributors
+  - <b> Fitting module </b>: Aniket Agarwalla, Nicolas Dupont, Carl Delrieu, Dorian Verge, Jules Vidal, Antoine Lafon, Anais Bains
+  - <b> SpatialPartitioning module </b>: Amael Marquez, Matthieu Gomiero, Dorian Ferreira, Sacha Vincent
 
   \subsection ponca_credits_subsection_grenaille Patate/Grenaille crew and contributors
   Ponca is a fork of Patate, module Grenaille, which is now deprecated, and which was developed by: