search

Teoge / DMPR-PS

DMPR-PS: A Novel Approach for Parking-Slot Detection Using Directional Marking-Point Regression
GNU General Public License v3.0
208 stars 48 forks source link

Abnormal model training #13

Open lzq04 opened 2 years ago

lzq04 commented 2 years ago

image

Follow your approach to PS2 0 data set, the training set and test set obtained, and the weight file obtained after training are used for parking point detection and evaluation. The evaluation data of parking point detection is shown in the figure below: image

In addition, the same data set is tested, and the weight file is modified to the author's. The evaluation data are as follows: image

The effect given in the original paper is

image So you need to explain to me why this happens? Or is there something wrong with my experiment?

lzq04 commented 2 years ago

The environment I use is pytorch1 7 and CUDA 10.2

Teoge commented 2 years ago

Average precision is a metric different from precision and recall. Take a look at https://en.wikipedia.org/wiki/Evaluationmeasures(information_retrieval)#Average_precision for details. If you wish to reproduce the precision recall in the paper, please refer to the step in here.