yzf0011的专栏

Monocular 3D Object Detection and Box Fitting Trained End-to-End Using Intersection-over-Union Loss insights 3DIou loss Homoscedastic noise Heteroscedastic noise 3D框的representation 重要结论 ...

M3D-RPN Post 3D–>2D Optimization 下图是M3D-RPN优化旋转角的方法 其优化角度的方式挺不错的,严格利用了3d框投影在图像上的8个角点会与2D框重合 该优化旋转角的方式,是否适合depth估计的迭代优化,即大致流程如下:d_init = predict() d_init_1 = d_init + sigma d_init_2 = d_init - sigm...