Revival of Complementary and Alternative Herbal Medicines

Herbal medicines, complementary or alternative medicines is a wide term for the therapies that are not part of standard care but it has many theories regarding efficacy based on personal experiences, history and common knowledge. It has long been used since ancient times since the beginning of human civilization. Its use had caught much attention in the early 1800s, with the development in the science of chemistry, a new era in pharmacotherapeutics and the use of active chemical ingredients in plants which were known to produce favorable therapeutic effects, were explored, active compounds were extracted, purified and their structure was revealed. This advancement paved the way towards modern pharmaceutical therapy. The modern drugs are based on these herbal medicines, after extracting the active and pure chemical compounds. Pharmacokinetics and physicochemical properties of the active ingredients was explored. It lead to the better understanding of efficacy and safety profile of these drugs and first choice for treatment of various diseases. At the same time, the herbal medicines were considered as secondarily important. After approximately two centuries, the use of herbal medicines have seen a revival globally both in developing as well as developed countries. In the past few years, the practice of using herbal medicines as an alternative and complementary health medicine has gained more importance. Herbal medicines are common for treatment of various ailments including cancer, digestive disorders, pain related disorders, neuropathic ailments and cardiac arrhythmias etc. Even it has been used by pregnant females and mostly perceived as safe. Its use has gained more attraction due to its ‘natural’ approach and lesser side effects. Their use if often overlooked but physicians should pay attention to these medicines. There is lack of familiarity, standardization of the drug components, unproven therapeutic effects in various diseases, unexplored toxicology, pharmacokinetics, drug-drug interactions, and compatibility in patients with varying medical, genetic and demographic history. There are serious concerns regarding the safety, efficacy and quality of herbal products and nutraceuticals. Accidental contamination and deliberate adulteration are assumed to be the main cause of the side effects. Much of the herbal medical knowledge is scattered in different regions of the world and mostly available at family, community and local level and mostly in any native languages. There is need of coherent sources, knowledge, and exploration of these medicines across the world. The herbal medicine has varying diversity in different geological regions and they should be investigated. There should be a regional or national body to control and approve the herbal medicines. Proper documentations on these medicines and food supplements should also be done.

Personalized Video Summarization Based on Viewers Emotions

Due to a rapid growth in the field of multimedia content, the user now demands video summaries, which represent the video content in a precise and compact manner according to their needs. Conventionally, video summaries have been produced by using a low-level image, audio and textual features, which are unaware of the viewer’s requirements and result in a semantic gap. Video content evokes certain emotions in a viewer, which can be measured and act as a strong source of information to generate summaries meeting viewer’s expectation. In this research, personalized video summarization framework is designed that classifies viewer’s emotion based on his/her facial expressions and electroencephalography (EEG) signals while watching a video to extract keyframes is presented. The first contribution of this thesis is to propose a new strategy to recognize facial expressions. For this purpose, the stationary wavelet transform is used to extract features for facial expression recognition due to its good localization characteristics, both in spectral and spatial domains. More specifically, a combination of horizontal and vertical sub-bands of the stationary wavelet transform is used as these sub-bands contain muscle movement information for the majority of the facial expressions. Feature dimensionality is reduced by applying discrete cosine transform on these sub bands. The selected features are then passed into a feed-forward neural network that is trained through back propagation algorithm to recognize facial expressions. The second contribution of this thesis is to generate personal video summaries with proposed facial expression recognition scheme. The video is shown to the viewer and facial expressions are recorded simultaneously using a Microsoft Kinect device. Those frames are selected as keyframes from the video, where different facial expressions of the viewer are recognized. The third and final contribution of this research is a new personalized video summarization technique based on human emotion classification using EEG signals. The video is shown to the viewer and electrical brain activity is recorded simultaneously using EEG electrodes. Features are extracted in time, frequency and wavelet domain to classify viewer’s emotion into happy, love, sad, anger, surprise and neutral. Those frames are selected as keyframes from the video, where the different emotions of the viewer are evoked. According to the experimental results the proposed facial expression recognition scheme using stationary wavelet transform gives an accuracy of 98.8%, 96.61% and 94.28% in case of Japanese Female Facial Expressions (JAFFE), Extended Cohn Kanade Dataset (CK+) and Microsoft- Kinect (MS-Kinect) datasets. Furthermore, it is evident from the results that the personalized video summarization using proposed facial expression recognition generates personal video summaries with high precision, recall, F-measure, accuracy rate, and low error rate, hence reducing the semantic gap. In case of emotion recognition using EEG signals, classification accuracy up to 92.83% is achieved by using support vector machine classifier when time, frequency and wavelet domain features are used in a hybrid manner. Experimental results also demonstrate that the proposed EEG based personal video summarization framework outperforms the state-of-the-art video summarization methods