Abstract

This research report is based upon the role of face detection approach and Artificial Intelligence (AI) in this new era of technology. This technology is referred to as one of the current emerging technologies and innovations which are approached. The overview of the research work and the details of the related work are being illustrated in this research report.

In this research the topic that will be discussed is of face detection using artificial intelligence. In the recent time the artificial intelligence has emerged a lot and it has be seen that the artificial intelligence is applied in most of the technology (Jindal & Kumar, 2013). By the application of the artificial intelligence the human being is able to lead a convenient life style. In the recent time the artificial intelligence is incorporated with many security devices and one of the new and innovative way is the face detection by using the artificial intelligence.

The face recognition is a technique which is capable to capture the pattern of the human face and store it for further usage. In this report the details will be given basis of the research in the face detection. The overview of the issues associated to Artificial Intelligence (AI) is also discussed in this paper. In the paper the discussion will be made on the technology and its application. The report will also state the research issues, challenges, related work, experimental analysis will be given. Finally a conclusion and recommendation will be made to complete the paper.

The concept of Artificial Intelligence needs more concentration while it is being used for detecting face. As the face detection is made mainly for the security purpose so there is a chance that the securities can be violated. Due to lack of time different issues associated to Artificial Intelligence (AI) are being overlooked but needs further discussion (Ghahramani, 2015). The face detection is mainly work as per the algorithm that is incorporated in the technology. There are many circumstances that the face detection tends to fail or might not work properly. So it is necessary to implement the face detection in such a way that the chance of the problems get low. The problems are basically general as it not that much important and day by day the technology is improving thus the issues and all the challenges associated with the face detection technology are getting reduced.

Research problem overview

In the recent time the face recognition is become an integral part for the all devices. The technology is used in many organisations and in many businesses (Ying et al., 2013). The face recognition is mainly used as the security equipment as it is assessed that a face detection system is more secure as every human has some unique facial data. The face recognition is used in the industries of manufacturing, law enforcements, in health care and in the construction.

As in the recent time the payment are made online so it is necessary to provide a secure way for the transaction. In the smart devices the face detection in applied. It is more secure and convenient than using the pass code or pattern. The face detection is also used in the computer to secure the sensitive data.

In the sector of law it stores the data of the criminals and if any record need to be find about a criminal then the face data are used (Kumar et al., 2017). The cameras that are AI equipped are using in many countries that helps to track human.

The artificial intelligence is able to control the computer and because of that the computers are able to think like the human being. The artificial intelligence mainly studies the way through which the human brains can read as well as learn from the external ambiance. Biometric scanning can be done through different ways such as face detection, iris scanner, fingerprint scanner etc. Thought the help of this approach the rate of criminal activities can be reduced. Artificial Intelligence (AI) is the framework through which data can be validated and verified at the same time. The face detection technology mainly stores the facial data of the human and the main part of the facial data contains the eyes (Li, et al., 2015). A logarithm is made for identifying the regions of the face and then the data are validated for the future use.

The face detection technology is also stated as the process which is psychological and can contain the visual scene of the face. The face detection technology is mainly regarded as the object class detection.

The matching process will not work if the image data is not matched with the data in the database (Zhang et al., 2013). The main use of the face detection is made on the biometrics. The artificial intelligence neural network is the core part of the face detection technology.

Applications of technology

With the many benefits that the face detection provides there are many issues and challenges that are associated with the face detection. The challenges with the face detection are given below:

Illumination variation: During face detection if the factors such as lighting (spectra, intensity, source distribution) are not utilized accordingly them serious issues may be faced by the researchers (Sermanet et al., 2013). This condition is one of the challenges of the face detection. The face data is captured under many illumination situation, if the light and the strength of the data differs then the face detection will not work properly.

Aging issues: If long time gap is being identified between the time of face detection and the time face matching then for certain changes the face may not be detected accordingly by the advanced forensic application.

Same face but if the scale of the face is different, then the face detection will have some issues. The distance from where the data is captured matters a lot.

In order to design the face recognition Artificial Intelligence (AI) mechanism, the concept of neural network is being utilized. The high dimensional space can be reduced into a very low dimension with the help of the PCA based eigenface approach. On the other hand, different successful methodologies are available that are widely being used by the developers in this previous era. One of the best coordinated systems used for the face detection is the “eigenface.” (Farfade, Saberian & Li, 2015) Besides this, different kinds of extraction methods such as linear discrimination analysis, kernel methods, evolutionary pursuit and support vector system etc for face images purposed in the past few years. Among all of these methods LDA is one of the supervised learning algorithms. The features of LDP are being obtained by computing the edge response values in all eight of the directions at each if the pixel positions. Evolutionary pursuit is another generic algorithm that is used to meet the purpose of face detection in the artificial intelligence. The process of human face detection is mostly identified as the first phase in any image that is prior to attempt the recognition which is much focused on the computational resources (Baltrušaitis, Robinson & Morency, 2016).  Another model named as MSNN model is a reliable which can again be utilized for the face detection purpose. The process has high level of accuracy as well as better success rates regardless of the working ambiance.

About technology

In order to avoid the challenges or issues of the human face detection a solution has been proposed which is related to the following phases:

• Adaboost combining artificial neural network for face detection
• Proper alignment of face using the concept of active shape model and multiple layered perceptions
• Feature extraction
• Matching of the features with the help of the multi artificial neural network

While making the face alignment using the perception of shape model and multilayered mechanism the detail activities that are being conducted:

• Building of a perfect shape model for the face alignment
• Usage of the ASM algorithm with the help of the 2D searching approach
• Multilayered perception modeling

On the other hand for the feature extraction the extract geometry face global as well as component feature vectors will be used (Jindal, N., & Kumar, 2013). In addition to this, ICA model will also be applied fo the successful accomplishment of the face recognition process.

The main issues associated to this technology are faced during the development of a trainable system for the face detection that is able to handle face rotation in depth and partially completely occluded faces. In order to overcome all of these challenges the detection system is required to be very much robust in nature. The process of face detection is referred to as the first phase of automated face recognition system (Ghahramani, 2015). In case of both the surveillance and human computer interface system this specific face detection mechanism is very much beneficial. In order to reduce the rate of criminal activities this face detection approach in terms of artificial intelligence is very much profitable. Different face detecting devices are also being designed and implemented that are being widely used by the developers. Automated surveillance wherever the objectives are to be recognized for tracing people which are on the watch list, needs effective concentration. In this open world of application all the systems are tasked for recognizing smaller set of people during the detection. There are many other potential areas for which the application is being widely used such as multimedia environment, airplane gates, sketch based, monitoring of close circuit television etc.

Conclusion

It has been found that, the automatic recognition of human faces is referred to as a significant issue in the development as well as application of the pattern recognition. From the overall discussion it can be concluded that in order to maintain the security of either the credentials or the financial records of any personnel or organization the biometric security approach in terms of Artificial Intelligence (AI) is very much helpful. Different biometric authentication approaches are available and based on the requirement of the organization the most suitable biometric authentication approach rather artificial intelligence technology should be adopted by the business organization or any personnel. Clustering is a very easy approach that is used by the security concern people for generating both the prototypes of face and non faced models. With the help of the support vector machine this face recognition process can be implemented. Again component based techniques can also be utilized for detecting the near frontal and frontal faces within the gray images using the SVM device. Face recognition is referred to as one of the fastest growing and challenge area in this real time application. The easiest and user friendly face detection algorithm sis elaborated on this research report.

Research issues and challenges

In order to resolve all the issues faced while using the concept of Artificial Intelligence in face detection the different recommendations those are useful mentioned in the below section:

Technology improvement: As this particular part of Artificial Intelligence mechanism is very much beneficial and currently using throughout thus one of the best possible solution for avoiding this error is to use high resolution cameras for the face detection that has better optical properties. Moreover around 120 pixels between the two eyes should be collected. In addition to this, this advanced approach will help to resolve the issues of poor optics as well as incorrect compression and focus.

Training and development: Professional training and development program should be arranged for those associates who are involved to this process of face detection to avoid human made errors. Again usage of automated image assessment is another approach that will help to reduce the function and operational issues.  

Privacy law: There are certain security level variables should be identified by the professionals such as privacy law, legislation and access control. It is defined that, these biometric approaches will resolve the issues of the Artificial Intelligence system.

References

Al-Allaf, O. N. (2014). Review of face detection systems based artificial neural networks algorithms. arXiv preprint arXiv:1404.1292.

Baltrušaitis, T., Robinson, P., & Morency, L. P. (2016, March). Openface: an open source facial behavior analysis toolkit. In Applications of Computer Vision (WACV), 2016 IEEE Winter Conference on (pp. 1-10). IEEE.

Farfade, S. S., Saberian, M. J., & Li, L. J. (2015, June). Multi-view face detection using deep convolutional neural networks. In Proceedings of the 5th ACM on International Conference on Multimedia Retrieval (pp. 643-650). ACM.

Farfade, S. S., Saberian, M. J., & Li, L. J. (2015, June). Multi-view face detection using deep convolutional neural networks. In Proceedings of the 5th ACM on International Conference on Multimedia Retrieval (pp. 643-650). ACM.

Ghahramani, Z. (2015). Probabilistic machine learning and artificial intelligence. Nature, 521(7553), 452.

Jaderberg, M., Simonyan, K., Vedaldi, A., & Zisserman, A. (2014). Synthetic data and artificial neural networks for natural scene text recognition. arXiv preprint arXiv:1406.2227.

Jindal, N., & Kumar, V. (2013). Enhanced face recognition algorithm using pca with artificial neural networks. International Journal of Advanced Research in Computer Science and Software Engineering, 3(6).

Koprinkova-Hristova, P., Mladenov, V., & Kasabov, N. K. (2015). Artificial Neural Networks. Eur. Urol, 40, 245.

Kumar, P. M., Gandhi, U., Varatharajan, R., Manogaran, G., Jidhesh, R., & Vadivel, T. (2017). Intelligent face recognition and navigation system using neural learning for smart security in Internet of Things. Cluster Computing, 1-12.

Li, H., Lin, Z., Shen, X., Brandt, J., & Hua, G. (2015). A convolutional neural network cascade for face detection. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (pp. 5325-5334).

Rautaray, S. S., & Agrawal, A. (2015). Vision based hand gesture recognition for human computer interaction: a survey. Artificial Intelligence Review, 43(1), 1-54.

Russell, S. J., & Norvig, P. (2016). Artificial intelligence: a modern approach. Malaysia; Pearson Education Limited,.

Sermanet, P., Eigen, D., Zhang, X., Mathieu, M., Fergus, R., & LeCun, Y. (2013). Overfeat: Integrated recognition, localization and detection using convolutional networks. arXiv preprint arXiv:1312.6229.

Xu, Y., Zhu, X., Li, Z., Liu, G., Lu, Y., & Liu, H. (2013). Using the original and ‘symmetrical face’training samples to perform representation based two-step face recognition. Pattern Recognition, 46(4), 1151-1158.

Ying, C., Qi-Guang, M., Jia-Chen, L., & Lin, G. (2013). Advance and prospects of AdaBoost algorithm. Acta Automatica Sinica, 39(6), 745-758.

Zhang, L., Jiang, M., Farid, D., & Hossain, M. A. (2013). Intelligent facial emotion recognition and semantic-based topic detection for a humanoid robot. Expert Systems with Applications, 40(13), 5160-5168.