HEAD POSE ESTIMATION is the main step of the face recognition process. Finding a method which can be fully automated and capable of responding to different POSEs in all aspects, has created a challenge for this topic. Compared to face detection and recognition, which have been the primary focus of face-related vision research, the HEAD POSE ESTIMATION is an inter processing step during the implementation of programs in machine vision systems. In this paper, a new algorithm is presented which first the face region is extracted from 3D HEAD image and then the feature vectors that strongly influence the human perception of HEAD POSE and these are extremely salient cues regarding the orientation of the HEAD are extracted. Finally, the extracted vectors are trained as input to the classifiers for comparing and differentiating features. In this paper, the Support vector Machines (SVM), the Radial Basis Function neural network classifier (RBF) and the K-nearest neighbor (KNN) methods were used to train feature vectors and data classification. The algorithm was tested on Frav3D and Gavab databases. Finally the experimental results of 98.48 percent of correct estimates which obtained from KNN method, indicate considerable enhancement compared to previous methods.

Background and Obejctives: Multi-task learning is a widespread mechanism to improve the learning of multiple objectives with a shared representation in one deep neural network. In multi-task learning, it is critical to determine how to combine the tasks loss functions. The straightforward way is to optimize the weighted linear sum of multiple objectives with equal weights. Despite some studies that have attempted to solve the realtime multi-person POSE ESTIMATION problem from a 2D image, major challenges still remain unresolved. Methods: The prevailing solutions are two-stream, learning two tasks simultaneously. They intrinsically use a multi-task learning approach for predicting the confidence maps of body parts and the part affinity fields to associate the parts to each other. They optimize the average of the two tasks loss functions, while the two tasks have different levels of difficulty and uncertainty. In this work, we overcome this problem by applying a multi-task objective that captures task-based uncertainties without any additional parameters. Since the estimated POSEs can be more certain, the proPOSEd method is called “CertainPOSE”. Results: Experiments are carried out on the COCO keypoints data sets. The results show that capturing the task-dependent uncertainty makes the training procedure faster and causes some improvements in human POSE ESTIMATION. Conclusion: The highlight advantage of our method is improving the realtime multi-person POSE ESTIMATION without increasing computational complexity.

With the increasing development of technology and the emergence of smart devices in today’s life, interactive communication with these smart devices has become inevitable. Among the many methods of interactive communication, which include the use of voice and clothes equipped with motion sensors, as well as image-based methods, each of which is focused on identifying and monitoring a specific part of the body, human hand-based methods due to their characteristics Especially the hand is of double importance in this field, and therefore the ESTIMATION of the three-dimensional state of the hand using image processing has become one of the most attractive and challenging research fields related to human-machine interaction. In this article, different methods of estimating the three-dimensional state of the hand are reviewed with an emphasis on methods based on image processing, and their strengths and weaknesses are stated and compared. This comparison is based on traditional methods as well as methods related to deep learning. Also, the databases used in hand posture ESTIMATION have been introduced and the characteristics of each of them have been investigated in different applications.

In the cases of severe paralysis in which the ability to control the body movements of a person is limited to the muscles around the eyes, eye movements or blinks are the only way for the person to communicate. Interfaces that assist in such communications often require special hardware or reliance on active infrared illumination. In this paper, we proPOSE a non-intrusive algorithm for eye gaze ESTIMATION that works with video input from an inexpensive camera and without special lighting. The main contribution of this paper is proposing a new geometrical model for eye region that only requires the image of one iris for gaze ESTIMATION. Essential parameters for this system are the best fitted ellipse of the iris and the pupil center. The algorithms used for both iris ellipse fitting and pupil center localization POSE no pre-assumptions on the HEAD POSE. All in all, the achievement of this paper is the robustness of the proPOSEd system to the HEAD POSE variations. The performance of the method has been evaluated on both synthetic and real images leading to errors of 2.12 and 3.48 degrees, respectively.

There are challenges such as depth perception and self-occlusion, in the field of 3D human POSE ESTIMATION and reconstruction which obstructs precise ESTIMATION of body joints. In this paper, we first extract human POSE by focusing on 2D ground-truth using sparse coding and. In the second approach, we use a learning-based Convolutional Neural Networks using sparse coding and a model based rectifier to extract the estimated POSE. POSE ESTIMATION by proPOSEdmethod has reduced the mean error of the reconstruction in comparison with the state of the artworks.

Automatic capture and analysis of human motion, based on images or video is important issue in computer vision due to the vast number of applications in animation, surveillance, biomechanics, Human Computer Interaction, entertainment and game industry. In these applications, it is clear that 3D human POSE ESTIMATION is an essential part. Therefore, its accuracy has a great effect on the performance of these applications. Because of the variation in appearance and articulations of human, self-occlusion and high dimensional state-space of human POSE, 3D human POSE ESTIMATION from image observations is a challenging problem. In this paper, a new method of 3D human POSE ESTIMATION from multi-view video sequence is introduced. In the proPOSEd method, instead of seeking directly over the high dimensional states-space of human POSE and employing the complex inferring algorithms, a hierarchical search method with distinct objective function for each part of the body and direct optimization methods is employed. Advantages of the proPOSEd method are: automatic initialization, labeling of parts of the body contour and using separate objective function for different parts of the body. Experimental results demonstrate that the proPOSEd method can be effectively used as a marker-less system to estimate 3D human POSE in a multi-view sequence.