vision4robotics / autotrack Goto Github PK

Thanks for your work, is it possible to give a demo code for AutoTrack？

引用了不存在的字段 'use_gpu'。

出错 get_colorspace (line 32)
if gparam.use_gpu

出错 get_features (line 51)
feature_pixels(:,:,currDim:(currDim+cf.fparams.nDim-1),:) = cf.getFeature(image,cf.fparams,gparams);

出错 AutoTrack_optimized (line 221)
x=get_features(pixels,features,global_feat_params);

出错 run_AutoTrack (line 87)
results = AutoTrack_optimized(params);

feature_pixels(:,:,currDim:(currDim+cf.fparams.nDim-1),:) = cf.getFeature(image,cf.fparams,gparams);找不到哪里使用了 get_colorspace函数

Hi, thank you for your good work.
I just want to ask in 4 uav benchmarks, you compare AutoTrack performance with some cf trackers, why do you compare with some siamese trackers, like siamrpn++ or other like DIMP50?

请问有python版本的吗

run_AutoTrack.m文件中第55行的参数

params.no_fram = seq.en_frame - seq.st_frame + 1;

en_frame-st_frame + 1指的是帧数吗

按照公式，正确的不该是这样吗？
% subproblem mu
if frame>2&&iter<admm_iterations
for i=1:size(g_f,3)
z(i)=power(norm(g_f(:,:,i)-g_pre(:,:,i),2),2)/(2*epsilon);
g_f_delta = g_f_delta + z(i);
end
mu=ref_mu-g_f_delta;
end