OpenCV determinant_1

This is determinant_0 with a different number on the end, and it behaves identically. determinant_1 also calls cv2.determinant - feed it a square matrix as an NPARRAY, get back its determinant as a float. The _0/_1 split is the opencv-comfyui generator's doing: OpenCV declares cv2.determinant twice in its Python stubs (once for MatLike, once for UMat), the generator makes a node per overload, and since both overloads collapse to the same NPARRAY type on the Comfy side, you get two functionally identical nodes. It's a quirk of the pack, not a feature you need to pick between. Most of the pack's families - dft_0/_1, dilate_0/_1, divide_0/1/2/3 - work the same way.

For the substance, see the determinant_0 page. The short version: the determinant is the number that tells you whether a matrix is invertible (nonzero = invertible, zero = singular) and something about its orientation. In image work it's genuinely relevant to geometric matrices - homographies and affine transforms - but ComfyUI's image-generation side rarely builds those, and this pack gives you no way to construct an arbitrary matrix anyway (its Image2Nparray/Nparrays2Image pair only round-trips actual images). So the realistic path to using this node is: build your matrix with a numpy node from another pack, wire it in, read the float out, and ask yourself whether you couldn't have just done np.linalg.det while you were already writing numpy.

Inputs: mtx (NPARRAY, square). Output: float - the determinant.

One honest use case worth keeping in your back pocket: sanity-checking a transform for degeneracy before feeding it somewhere that will divide by it. A determinant near zero means your transform is about to do something catastrophic to pixel coordinates. That's a legit debugging workflow. It's just niche.

Install: identical to the rest of the pack - it ships in geroldmeisinger/opencv-comfyui, install via ComfyUI Manager (search "OpenCV") or:

cd ComfyUI/custom_nodes
git clone https://github.com/geroldmeisinger/opencv-comfyui

restart, and confirm opencv-contrib-python is present (pip install opencv-contrib-python). If the whole pack fails to load with Cannot import name 'guidedFilter' from 'cv2.ximgproc', you've got conflicting OpenCV wheels - resolve that first; it's the pack's most common install failure.

If you're deciding between _0 and _1, you're overthinking it. They're the same node. Pick one.