Depth-based face recognition using local quantized patterns adapted for range data

Abstract

A depth-based face recognition algorithm specially adapted to high range resolution data acquired by the new Microsoft Kinect 2 sensor is presented. A novel descriptor called Depth Local Quantized Pattern descriptor has been designed to make use of the extended range resolution of the new sensor. This descriptor is a substantial modification of the popular Local Binary Pattern algorithm. One of the main contributions is the introduction of a quantification step, increasing its capacity to distinguish different depth patterns. The proposed descriptor has been used to train and test a Support Vector Machine classifier, which has proven to be able to accurately recognize different people faces from a wide range of poses. In addition, a new depth-based face database acquired by the new Kinect 2 sensor have been created and made public to evaluate the proposed face recognition system.

Hand-gesture-based human-machine interface system using Compressive Sensing

“…The key contribution of the paper is the introduction of a CS stage to reduce the dimensionality of Depth Local Quantize Pattern (DLQP) feature descriptor [5], which are highly discriminative, but inoperative to use without any reduction because of their high dimensionality. Unlike other reduction techniques, CS is able to largely reduce the dimension of the descriptor while preserving the most of the information, since the relative distances among the reduced feature descriptors are (almost) preserved.…”

RGB-D datasets using microsoft kinect or similar sensors: a survey

“…Depth images (512×424 pixels) are saved with 16 bits format. One recent published paper about HRRFaceD can be found in [62]. Sample images from this dataset are shown in Fig.…”

“…Therefore, the scene recognition rates are relatively low, which are 75.9±2.9 (RGB), 65.8±2.7 (Depth) and 76.2±3.2 (RGB-D) reported in [42]. The latest algorithm performance comparisons based on B3DO, Person Re-identification, Kinect FaceDB, Big BIRD and HRRFaceD can be found in [32,40,64,66] and [62] respectively.…”

“…The image processing group in Polytechnic University of Madrid created high resolution range based face dataset (HRRFaceD) [62] in 2014, intending to evaluate the recognition of different faces from a wide range of poses. This dataset was recorded by the second generation of Microsoft Kinect sensor.…”

Using nonlocal filtering and feature extraction approaches in three-dimensional face recognition by Kinect

“…Recently, a newer version of the device has been introduced, which has both higher color and depth resolution, wider field of view, and can sense farther objects more accurately. 11 There are some studies published recently which used Kinect for face recognition, 10,[12][13][14][15][16] although the reported performance of them is acceptable, but they mostly emphasize on the different methods of extracting features from RGBD images, but it seems that the main issue in the usage of Kinect for recognition tasks is reducing the high level of noise in captured images by Kinect.…”