Clarify problem 5.2 a (#362)

book/clutter/exercises/normal_estimation_depth.ipynb

@@ -248,7 +248,7 @@
     "## Computing Surface Normals by Nearest Pixels\n",
     "Now we should be able to calculate the surface normals. Recall from section 5.5.2 of the textbook, in each sliding window we can use the points to construct a data matrix (also known as a *scatter matrix*) **W** which exhibits special properties that allows us to estimate the local normals. \n",
     "\n",
-    "**Problem 5.2.a** [2pts] Which eigenvector of **W** corresponds to the vector normal to the points in the sliding window? Assume we want the normal to align with the camera's z-axis, how can you use the eigenvectors of **W** to create a rotation matrix representing the desired orientation of the normal frame. Justify that your answer represents a valid rotation matrix. "
+    "**Problem 5.2.a** [2pts] Which eigenvector of **W** corresponds to the vector normal to the points in the sliding window? Assume we want to specify a normal frame whose z-axis corresponds to the vector normal to the points in the sliding window. How can you use the eigenvectors of **W** to create a rotation matrix representing the desired orientation of the normal frame. Justify that your answer represents a valid rotation matrix."
    ]
   },
   {