SiftGPU is an implementation of SIFT [1] for GPU. SiftGPU processes pixels parallely to build Gaussi

源代码在线查看: instruction.txt

软件大小: 6821 K
上传用户: autojacky
关键词: SiftGPU implementation processes parallely
下载地址: 免注册下载 普通下载 VIP

相关代码

				
				The list of required libraries is 
				
				libCg.so	libCgGL.so	libGLEW.a	libglut.so
				libIL.so -> libIL.so.1 		libIL.so.1	libCudart.so
				
				Those libraris may also be dependent on other libraries, and you may need to download them yourself. You also need to handle the compatibility of those libraies yourself. 
				
				Get those libraries either in the system default path or ./bin here before compiling. 
				
				The output binaries are 
				./bin/libsiftgpu.a	./bin/TestWinGlut	./bin/SimpleSIFT 	./bin/speed
				
				
				If you want to run the stuff with any .so files in ./bin, don't forget to add . to LD_LIBRARY_PATH by calling:
				export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:.
				
				
				If some of your header files (e.g. GL/gl.h )are not in the default path (e.g. /usr/include ), you'll need to add extra flags to CFLAGS in the makefile. Say, your opengl header are in /myheaders/GL. You'll need to add -isystem/myheader to CFLAGS.  			

相关资源