A novel Human Detection approach Based on Depth Map via Kinect

In this paper, a new method of human detection based on depth maps from 3D sensor Kinect is proposed. First, the pixel filtering and context filtering is employed to roughly repair defects on the depth map due to Informatio n inaccuracy captured by Kinect. Second, a dataset consisting of depth maps with various indoor human poses are constructed as benchmark. Finally, by introducing Kirsch mask and three-value codes to local Binary pattern, a novel local ternary Direction pattern (LTDP) feature descriptor is extracted and was used for human detection with SVM as classifier. The performance for the proposed approach are evaluated and compared with other five existing feature descriptors using the Same SVM classifier. Experiment results manifest the effectiveness of the proposed approach.

Man's detection

These methods can is roughly divided into three different categories; Human model based methods [1], template matching based methods [2] and statistical classification methods [3-5].

LBP (local binary pattern) feature, which is a string of bits obtained by binarizing Local neighborhood of pixels with Res Pect to the brightness of central pixel, were recently proposed to capture microscopic local image texture and was applied For human detection successfully

Hog-lbp

LTP (Local ternary patterns)

Centrist (census transform of histograms)

Unfortunately, since the Kinect mainly depends on speckle method [+], the depth map captured by Kinect often contains MUC H noise. So, Kinect uses this depth image to detect inaccurate, unstable

Overview of lbp-related Features

By defining the number of spatial transitions (0/1 changes) in LBP pattern with a U value defined as below in (3), the Uni Formity of LBP patterns, which refers to the patterns have limited transition or discontinuities (U2) in the circular bi Nary presentation, can be determined, where the U value is given as

The uniform LBP only have a bins, one for possible uniform patterns and one for all of the non-uniform ones.

Ternary LTP Code

In summary, the UNIFORM-LBP reduces the dimensions of LBP, while LTP extends LBP to three-valued codes and therefore Enhan Ces its anti-noise performance. The centrist introduces a pyramid structure to LBP and makes a multi-scale observation.

Feature Extraction

Noise Reduction Filters to Depth Map

Compared with TOF data, the depth map captured by Kinect have mounts of null-value areas, which present as ' white holes ' in Depth map

Traditional filters, E.I, mean filtering, Gaussian filtering, usually is utilized to remove salt and pepper noises

The pixel filter is designed to compensate the ' holes ' and the context filter is employed to further reduce noise in GE Neral.

It should is noted that the frames waiting to be retrieved is limited, because the method ignores the scene change Betwee N Frames.

The LTDP Feature

LTDP (local ternary direction pattern) is derived from the lbp-related feature descriptors by plugging a Specific Kirsch mask [] and three-valued codes to it

At first, the LTDP was calculated by comparing the relativ E Edge response value of a pixel in different directions. The Edge response value (S0~S7)   of a particular pixel is calculated with the Kirsch mask at eight different directio Ns. The Masks (M0~M7) is shown in fig.3.

Second, the 3-valued codes for the eight directions with threshold T is defined as follows

I n the experiment of this paper, a uniform pattern argument are designed and a coding scheme is used to split each ternary p Attern to its positive and negative halves and subsequently treating them as-separate channels of LTDP features for Which separate histograms is computed by combining the results is at the end of the computation.

Histogram of LTDP Feature for Depth Map

A depth map should is divided into several non-overlapping rectangular blocks.

A spatial histogram, concatenating the histograms of all blocks can is employed to represent the whole image.

Classification algorithm

SVM is relatively-robust and easy-to-be implemented. In this study, both linear kernel SVM and nonlinear kernel SVM is used as classification algorithm.

Detection class Read the first paper, also pointed out the idea of face detection bar ~ ~ ~

A novel Human Detection approach Based on Depth Map via Kinect