Description:FIELD OF THE INVENTION
Our Invention is related to a Intelligent 3D human sensing using machine learning

BACKGROUND OF THE INVENTION

Maintaining organisational effectiveness and enhancing effective iterations of organisational change at the corporate level both need an effective human resource management (HRM) strategy.

In order to lighten the strain and minimize HRM expenditures, the majority of HRM teams are presently integrating intelligent management technologies into their organization's people management activities.

In the field of intelligent HRM, several researchers have put forth numerous study methodologies, and numerous research findings have been made. The sophisticated HR management system has a sizable system.

We will begin the research on the HR punch card assist system in light of the multifunctionality and complexity of the HR system as well as our recent findings. The face recognition clock-in system is the one that has been used the most in the clock-in assistance system.
OBJECTIVES OF THE INVENTION
1) The objective of the invention is to provide a “Intelligent 3D human sensing using machine learning” is a Enterprises may increase their organisational and operational efficiency by using digital human resource management.
2) The other objective of the invention is to provide a 3D face recognition into the digital human resource management system to increase the effectiveness of corporate digital management and address issues with the low security level and unstable 2D face recognition.
3) The other objective of the invention is to provide a develop a digital face recognition management system and provide a face identification technique based on a multistream convolutional neural network and local binary pattern. We initially create the computer vision scenario for the system.
4) The other objective of the invention is to provide a depth camera image extraction approach is then used to create a local binary mode facial expression feature extraction system.
5) The other objective of the invention is to provide a construct a multistream convolutional neural network to learn facial 3D characteristics since it is simple to overlook facial 3D features.

SUMMARY OF THE INVENTION
We have developed a three-dimensional space-based face recognition network and rebuilt the face recognition technique. In order to prevent frame loss, we first construct a digital computer vision scenario using three REALSENSE depth cameras. Each depth camera's RGB image sensor and depth infrared sensor are linked together.

The physical feature calculation unit is initially used to combine the depth information and RGB information to produce 3D facial information. The facial RGB image information and depth information are concurrently fed to the face recognition model.

Two sets of 3D facial contour characteristics and data stream face categorization theme scores make up the 3D facial information. A unique face estimation technique extracts the characteristics from the RGB picture data, which has a significant weight in the predicted score.

The weight parameters for the score estimate are corrected using the various levels of features in the database, which are produced through repeated iterations.

BRIEF DESCRIPTION OF THE DIAGRAM

Fig.1: Intelligent 3D human sensing using machine learning Flow.
Fig.2: Intelligent 3D human sensing using machine learning Process
Fig.3: Intelligent 3D human sensing using machine learning method.

DESCRIPTION OF THE INVENTION

The digital face recognition system is comprised of a computer vision unit, a deep convolutional neural network algorithm unit, a GPU-accelerated computation unit, a data storage unit, and a PC control unit, in accordance with the project requirements.

With cameras positioned at 45 degrees and kept at the same level, we employ the stereo system architecture platform AVT Pike in the computer vision unit construction to aid in the capture of facial image and depth information.

Camera No. 1 is pointing in the direction of the subject's face, while cameras Nos. 2 and 3 are 45 degrees apart from camera No. 1. There are 1000 mm between the cameras and the face.

The combined field of view of the three cameras is 500 mm. The three cameras' lenses are connected to the main console using IEEE_1394 firewire. Figure 6 displays the specifics of the computer vision system construction. Tensor is used by one of the deep convolutional neural network modules.

Python serves as the language for altering algorithms and serves as the fundamental network infrastructure. The computer's 64 GB of RAM, Intel(R) CoreTM i9-9900 K CPU@3.6 GHz x8.
NVIDIA Quadro P6000 graphics card are used to train facial recognition models. To optimise the training parameters, the network training method employs a hierarchical training model with layer-by-layer update rounds.

, Claims:I/WE CLAIMS

1. Our Invention “Intelligent 3D human sensing using machine learning” is a Enterprises may increase their organisational and operational efficiency by using digital human resource management. We integrate 3D face recognition into the digital human resource management system to increase the effectiveness of corporate digital management and address issues with the low security level and unstable 2D face recognition. We develop a digital face recognition management system and provide a face identification technique based on a multistream convolutional neural network and local binary pattern. We initially create the computer vision scenario for the system. The depth camera image extraction approach is then used to create a local binary mode facial expression feature extraction system. We construct a multistream convolutional neural network to learn facial 3D characteristics since it is simple to overlook facial 3D features.
2. According to claim1# the invention is to a “Intelligent 3D human sensing using machine learning” is a Enterprises may increase their organisational and operational efficiency by using digital human resource management.
3. According to claim1,2# the invention is to a 3D face recognition into the digital human resource management system to increase the effectiveness of corporate digital management and address issues with the low security level and unstable 2D face recognition.
4. According to claim1,2,3# the invention is to a develop a digital face recognition management system and provide a face identification technique based on a multistream convolutional neural network and local binary pattern. We initially create the computer vision scenario for the system.
5. According to claim1,2,4# the invention is to a depth camera image extraction approach is then used to create a local binary mode facial expression feature extraction system.
6. According to claim1,3,5# the invention is to a construct a multistream convolutional neural network to learn facial 3D characteristics since it is simple to overlook facial 3D features.