Help about XNOR-Popcount on FPGA/Verilog

[Starchasm Nyx]

Hello,I have a final project about edge detection with FPGA/Verilog.But instead of using normal convolution with kernel like Sobel,my teacher asked me to use XNOR-Popcount with pre-defined kernel/mask.
I want to ask that is XNOR-Popcount can replace Convolution with Sobel for edge detection? And if that's possible,can i use Sobel as pre-defined kernel/mask for my project?
Thank everyone and i hope someone can help me.

 

[FlyingDutch]

Hello,

some time ago I tried project wich has used "Sobel Filter" on Xilinx Artix7 FPGA. You can use "Vivado HLS" (High Level Synthesis" to convert algorithms written in C/C++ languages to HDL versions (VHDL or Verilog). I don't know if this be helpful for you, but I am pasting link to polish forum (Translated by "Google Translate" into English):

https://forbot-pl.translate.goog/fo...=en&_x_tr_hl=pl&_x_tr_pto=wapp#comment-151126

Best Regards

 

[Starchasm Nyx]

Hello,

some time ago I tried project wich has used "Sobel Filter" on Xilinx Artix7 FPGA. You can use "Vivado HLS" (High Level Synthesis" to convert algorithms written in C/C++ languages to HDL versions (VHDL or Verilog). I don't know if this be helpful for you, but I am pasting link to polish forum (Translated by "Google Translate" into English):

https://forbot-pl.translate.goog/fo...=en&_x_tr_hl=pl&_x_tr_pto=wapp#comment-151126

Best Regards

Thank for your reply
I have done a project with sobel already.But in my project i've used convolution matrix input with sobel.But now i must use xnor-popcount to replace convolution.So i dont know if possible xnor-popcount can replace conv and which mask should i use for xnor

 

[FlyingDutch]

Thank for your reply
I have done a project with sobel already.But in my project i've used convolution matrix input with sobel.But now i must use xnor-popcount to replace convolution.So i dont know if possible xnor-popcount can replace conv and which mask should i use for xnor

Hello,
unfortunately I didn't use "Binary Convolution Neural Networks" so I am not able to help.

Best regards