2021
DOI: 10.1109/tgrs.2020.3010880
|View full text |Cite
|
Sign up to set email alerts
|

Multidimensional Feature Representation and Learning for Robust Hand-Gesture Recognition on Commercial Millimeter-Wave Radar

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1
1

Citation Types

0
6
0

Year Published

2021
2021
2023
2023

Publication Types

Select...
6
1
1

Relationship

0
8

Authors

Journals

citations
Cited by 56 publications
(15 citation statements)
references
References 41 publications
0
6
0
Order By: Relevance
“…17,18 Therefore, only by integrating the images and data of active and passive millimeter wave imaging, can target objects with different shapes and materials be identified more accurately. According to the image feature selection criteria, [19][20][21] the following image feature extraction is studied in this paper.…”
Section: Image Feature Extraction and Object Classification Recognitionmentioning
confidence: 99%
“…17,18 Therefore, only by integrating the images and data of active and passive millimeter wave imaging, can target objects with different shapes and materials be identified more accurately. According to the image feature selection criteria, [19][20][21] the following image feature extraction is studied in this paper.…”
Section: Image Feature Extraction and Object Classification Recognitionmentioning
confidence: 99%
“…Compared with the IR-UWB radar, the FMCW radar provides multidimensional information including target range, velocity and angle with multiple antennas. The FMCW signal of the i-th detected people with a 2-D MIMO antenna array is expressed as follows [23]:…”
Section: Fmcw and Ir-uwb Radar Signal Modelmentioning
confidence: 99%
“…The RDM was then averaged over time to obtain RD(r, v) for target detection. The neighbor threshold detection method [23] was applied iteratively to select the point with the local maximal amplitude in RD(r, v), detecting the distance r i and velocity v i of the i-th target. The azimuth angle α i was then computed with angle FFT.…”
Section: Target Detection and Roi Selectionmentioning
confidence: 99%
“…Gesture recognition has been implemented with multiple types of sensors, such as cameras [7,8], and radars [9] or their combination [10]. Most of these works explore obtaining highly discriminative features, for example, Ge et al used a 3D-CNN-based method for 3D volumetric representation of the hand depth image to obtain 3D spatial information [8], and Molchanov et al combined radar, image, and depth sensors to provide complementary information about the shape, color, and the instantaneous angular velocity of the hand [10].…”
Section: Motivationmentioning
confidence: 99%