Vincent Rabaud comments

Results 89 comments of


                                            Vincent Rabaud

Get CUDA code to compile with clang CUDA and without CUDA

@asmorkalov , for my purpose, I do not use nvcc but clang cuda instead. This is all internal to Google so I actually have no idea of the command lines...

Get CUDA code to compile with clang CUDA and without CUDA

I am done adding missing implementations, PTAL

VideoCapture add support for byte buffer

May I suggest to actually support `vector` (or just raw pointer) instead of cv::Mat to not fall into the problem with int/size_t like in https://github.com/opencv/opencv/issues/22036

Integrate gpb-ucm from https://github.com/vrabaud/gPb and add python binding

@alalek , that was a GSOC and Jitendra Malik gave the rights to the OpenCV foundation (we can talk about it offline with Vadim if you want)

Changes about when APNG has a hidden frame

The alpha compositing is tricky when the alphas are both non full. https://ciechanow.ski/alpha-compositing/ expiains it well. Your formula in uint16_t and uint8_t should be the same.

isContinuous() fails for large images in OpenCV

@asmorkalov , @alalek should that renaming be done for OpenCV 5 ?

isContinuous() fails for large images in OpenCV

@vpisarev , I was just mentioning the renaming from `isContinuous` to `isContiguous` in https://github.com/opencv/opencv/issues/12018#issuecomment-408563033 , as OpenCV 5 can break the API. Not sure it is worth it, I just...

Fix expression result error [-Werror,-Wunused-value] when compiling with clang

Until C++17 and "maybe_unused", for now, you can also just cast to (void), that is the usual convention.

Fix expression result error [-Werror,-Wunused-value] when compiling with clang

LGTM

Corrected -Wshift-sign-overflow warnings to prevent negative results

Can you please replace the "U" with "ll" (for long long, cf https://en.cppreference.com/w/cpp/language/integer_literal.html) instead to go to 64 bits? That should solve most problems.