Gary Chan

 + **The Green Boxes** are the segmentation result after tracking help. + **The Dark Green Boxes** are the internal tracking-help predicted result. ### Could we keep...

## How to download? 1. Search Dataset on [KiTTI's Raw Data](http://www.cvlibs.net/datasets/kitti/raw_data.php) 2. Recommend to download using [XunLei](https://www.xunlei.com/), for example  + [2011_09_26_drive_0005's synced+rectified data](https://s3.eu-central-1.amazonaws.com/avg-kitti/raw_data/2011_09_26_drive_0005/2011_09_26_drive_0005_sync.zip) + [2011_09_26_drive_0005's calibration](https://s3.eu-central-1.amazonaws.com/avg-kitti/raw_data/2011_09_26_calib.zip) + [2011_09_26_drive_0005's tracklets](https://s3.eu-central-1.amazonaws.com/avg-kitti/raw_data/2011_09_26_drive_0005/2011_09_26_drive_0005_tracklets.zip)

 ## Coordinate Systems 


## [CSDN: KITTI数据集简介与使用](https://blog.csdn.net/solomon1558/article/details/70173223)

> Teichman A, Levinson J, Thrun S. Towards 3D object recognition via classification of arbitrary object tracks[C]//Robotics and Automation (ICRA), 2011 IEEE International Conference on. IEEE, 2011: 4034-4041.  

## 计算precision >  ## 计算recall > 

## [AP计算由来](https://github.com/bostondiditeam/kitti/wiki/Evaluation-Metric)  + Let us understand how all this makes sense. Lets say you built an algorithm that is extremely efficient at detecting the back of the cars. Such...

## [AP 计算公式](https://blog.csdn.net/weixin_35653315/article/details/79669601)  ```cpp vector getThresholds(vector& v, double n_groundtruth) { // holds scores needed to compute N_SAMPLE_PTS recall values vector t; // sort scores in descending order // (highest...

## 借助Ground Truth's IoU Overlap计算TP、FP > For `cars` we require an 3D bounding box overlap of **70%**, while for `pedestrians` and `cyclists` we require a 3D bounding box overlap of...