(An ISO 9001:2008 Certified Online Journal) ISSN:2455-9660


Volume 01 Issue 09 (November-2016) | IJERAS

Title: Diminishing Public Health Due to Particulate Matter in the Ambient

Authors: Charu Khosla Gupta and Arijit Chowdhuri

For Abstract: Click Here
Human beings have made a noteworthy progress in terms of technology, industry, generation of different forms of energy, fuels, fertilisers etc, which has led to enormous amounts of harmful pollutants in the ambient. According to the latest urban air quality database, 98% of cities in low- and middle income countries with more than 100 000 inhabitants do not meet WHO air quality guidelines. Particulate Matter is the deadliest of all primary air pollutants kindling larger health concerns because continuous exposure to it leads to increased risk of cardio-respiratory morbidity, mortality, lung cancer and reduction in life expectancy. The current paper is a review on composition and sources of particulate matter and their subsequent affects on public health and well being. To conclude, a few measures for mitigation are also discussed.

Page 1-5


Title: Analyzing the Use of Partial Coefficients of Hybrid Orthogonal Transforms for CBVR

Authors: Dr. Dhirendra S. Mishra, Ms. Pallavi R. Rege

For Abstract: Click Here
Content based video retrieval (CBVR) systems use the contents of video for identifying and representing videos for retrieval purpose. Video data are large in volume and complex in nature. Therefore, CVBR needs compact representative feature vectors for content of stored and query videos such that comparisons complexity is reduced. Partial coefficients of orthogonal transforms can be used as feature vector for retrieval [1-3]. In this paper the partial coefficients of seven hybrids, using kronecker product [4], of seven orthogonal transforms are used for CBVR. Four blocks sizes, 6x6, 8x8, 12x12, 16x16, of coefficient are considered. Four different similarity measures are used and results are analyzed using performance measure based on precision and recall. The dataset includes 300 videos in 10 categories which is queried by query videos. It was observed that hybrids of discrete cosine transform performed best when cosine correlation measure was used.

Page 6-9