CS2125 Paper Review Form - Winter 2019

Reviewer: Abdul Kawsar Tushar

Paper Title: Springrobot: A Prototype Autonomous Vehicle and Its Algorithms for Lane Detection

Author(s): Qing Li, Nanning Zheng, and Hong Cheng

1) Is the paper technically correct?
 [X] Yes
 [ ] Mostly (minor flaws, but mostly solid)
 [ ] No

2) Originality
 [X] Very good (very novel, trailblazing work)
 [ ] Good 
 [ ] Marginal (very incremental)
 [ ] Poor (little or nothing that is new)

3) Technical Depth
 [X] Very good (comparable to best conference papers)
 [ ] Good (comparable to typical conference papers)
 [ ] Marginal depth
 [ ] Little or no depth

4) Impact/Significance
 [ ] Very significant
 [X] Significant
 [ ] Marginal significance.
 [ ] Little or no significance.

5) Presentation
 [ ] Very well written
 [X] Generally well written
 [ ] Readable
 [ ] Needs considerable work
 [ ] Unacceptably bad

6) Overall Rating
 [X] Strong accept (award quality)
 [ ] Accept (high quality - would argue for acceptance)
 [ ] Weak Accept (borderline, but lean towards acceptance)
 [ ] Weak Reject (not sure why this paper was published)


7) Summary of the paper's main contribution and rationale
   for your recommendation. (1-2 paragraphs)

This paper proposes a novel algorithm of lane detection for vehicles, which paves the way of usage of this algorithm in autonomous vehicles. The base of this work is the proposed adaptive randomized Hough transform method. 

A thing that concerns me is how much the takeaway is in the present situation, since there are novel algorithms with better performance in the present research domain.

Other researchers have cited this paper for its early work on the lane detection system.

8) List 1-3 strengths of the paper.  (1-2 sentences each,
identified as S1, S2, S3.)

S1. Very novel work in the contemporary setting.

S2. Good technical depth.


9) List 1-3 weaknesses of the paper (1-2 sentences each,
identified as W1, W2, W3.)

W1. The image processing and corner detection techniques were popular once up a time. After the advent of deep learning and specifically Convolutional Neural Netowrks (CNNs), these seem to be the most popular algorithms for any computer vision problem. Therefore, although this proposes a novel and influencing algorithm for the contemporary research world, it probably has lost quite a bit of relevance by now.

W2. The paper has only worked with R and G channels of an RGB image by virtue of them being consistent with the whiet and yellow lane markings. However, where the lane markings are discolored and hence the original color of the road is exposed, this could be fatal. In any way, not taking advantage of all the information of and image does not seem particularly clever. 

W3. This paper only works with images, not videos. There are many recent technologies that can work with videos, such as YOLO.