tj27-vkr / rcnn-vehicle-tracking-lane-detection Goto Github PK

link for the pre trained weights already expires

@tj27-vkr Hi, I am amazed of you reimplementation of Mask-RCNN with the land detection branch and road segmentation branch added. So, how have you trained your mask rcnn project? especially what's your training dataset?

I was able to get everything installed and working wih included test images (e.g. test6.jpg). However, when I used my own input images and video, the lane detection did not work very well. I wonder if this issue is related to image size??? Both the video and images were 1280x720.

Code:
https://pastebin.com/FEuxqVup

Dear author:
I am new learner for programming. I want to implement the project, but I don not know what software environment this project needs. Now , I installed python and other modules by anaconda .When I run main.py, there's a wrong hint. Could you tell me what software and hardware environment I need. Do I need GPU?

Are there any requirements about input image?because when I use my own image to test ,there is error

Hi,

I am going to use your algorithm for a simulation, but when I replaced a produced image from the simulator with your images, the detected lines aren't so accurate! As you can see in the attached picture the left line isn't matched with the yellow lane! Would you please tell me how can I increase the accuracy of your algorithm?

from keras.preprocessing import image
this line is showing error and the pre-installed dataset i am not able to connect the image path

Hello, can I use my dataset to train this project and how to do it?