Suppose we have $f(x,y) = x^4+y^4+2x^3$, and trying to find the extreme values of the function. We have \begin{align*} & D_1f(x,y) = 4x^3+6x^2, \\ & D_2f(x,y) = 4y^3. \end{align*} and solving the following system \begin{align} 4x^3+6x^2 = 0, \\ 4y^3 = 0. \end{align} (1) gives $x^2(4x+6) = 0$, so $x = 0$ or $x = -\frac{3}{2}$. (2) gives $y = 0$. So we find two stationary points: $(0,0)$ en $(-\frac{3}{2},0)$.
For the Hessian we have \begin{align*} & D^2_1f(x,y) = 12x(x+1), \\ & D_{12}f(x,y) = 0, \\ & D^2_2(x,y) = 12y^2. \end{align*}
We note that $(0,0)$ is a saddle point. We have $f(0,0) = 0$. For $\epsilon \in (0,1)$ it holds that $f(-\epsilon,0) = (-\epsilon)^4 + 2(-\epsilon)^3 = \epsilon^4 -2\epsilon^3 < 0$, but on the other hand that $f(0,\epsilon) = \epsilon^4 > 0$. So $(0,0)$ is a saddle point.
Now I am supposed to prove that $\vec{x}_1 = (-\frac{3}{2},0)$ is a local minimum using level curves, i.e. $f(x,y) = c$. Note that $f(-\frac{3}{2},0) = -\frac{27}{16}$. The hessian is given by $$ H_f(-3/2,0) = \begin{bmatrix} 9 & 0 \\ 0 & 0 \end{bmatrix} $$ which is inconclusive, since it has eigenvalues $0$ and $9$. However, it follows that $\vec{x}_1$ cannot be a maximum since it is positive semidefinite.
For a level curve we have $x^4+y^4+2x^3 = c$, whence $y = \sqrt[4]{c-x^4-2x^3}$. So the level curve of level $c$ is parametrised by the curve $g(t) = (t, \sqrt[4]{c-t^4-2t^3})$, with $c \geq t^4-2t^3$.
How do I proceed from here to prove that $\vec{x}$ is a local minimum?
Taking
$$ f(x,y) = x^4+y^4+2x^3 $$
and making a coordinates change as
$$ x^2 = u\\ y^2 = v $$
the new function reads
$$ g(u,v) = u^2+v^2+2u\sqrt u $$
and the Hessian is
$$ H_g = \left( \begin{array}{cc} 2+\frac{3}{2 \sqrt{u}} & 0 \\ 0 & 2 \\ \end{array} \right) $$
We can observe that at $x= \epsilon$ we have
$$ H_f = \left( \begin{array}{cc} 2+\frac{3}{2 \epsilon^2} & 0 \\ 0 & 2 \\ \end{array} \right) $$
which is definite positive