kumoh national institute of technology
Networked Systems Lab.

Ali Aouto, Thien Huynh-The, Jae-Min Lee and Dong-Seong Kim, "Pose-Based Identification Using Deep Learning for Military Surveillance Systems", The 10th International Conference on ICT Convergence October 2019
By :
Date : 2019-09-04
Views : 176

Review 1 : reject

The paper states it proposes a gait recognition scheme based on spatiotemporal poses on the abstract and a military surveillance system on the conclusion. The authors should make it clear what is the real contribution of their work.
Few other points also have to be considered:
- Text is well written, but it has some spelling mistakes.
- The authors should explain better the experimental scenario. It is not clear how they trained and tested.
- Likewise, the results are not explained.
- Authors might also consider citing the algorithms they used.

Review 2 : neutral

The authors proposed a gait recognition scheme utilizing the skeletal data extracted from Kinect cameras. In their proposed method, features of joint pairs vectors are used. The features compose of distance between joints and angles between the vector and the coordinate axis. After that they feed the input features to some pre-trained CNN networks to get the corresponding outputs. In the experiments, they only compare the results between some networks applied and do not compare with the referenced researches. There are some points, the authors should notice: 1) The referenced papers (where the authors use the same gait dataset) have better accuracy. 2) The paper is not well written and not clear about the identification implementation details with the chosen dataset. 3) There are typos in both sentences and mathematical formulas (e.g. (1)). 4) The title and abstract of the paper are not dealt with in the experiments.

Review 3 : Accept

no comment