ESTIRJ Volume 5, Issue3  Sept, 2021

Paper 1: AN APPLICATION OF HUMAN FALL PREDICTION USING CONVOLUTIONAL NEURAL NETWORK (1-6)

Abstract: Falls are major problem especially in elderly people, that causes major health problems. It became a source of deaths, internal injuries, and a loss of anatomy. Therefore, comprehensive research is required in predicting the falls, on the right time. Or Timely prediction of falls can minimize the death or injury rate as well. Moreover, it is shown that most of the falls detection methods having deep learning methodologies. And in our research, we used most effective and useful deep learning method for fall detection that is CNN (Convolutional Neural Network) with two publicly available datasets that is URFD. Our experimental results shows that CNN performed well in terms of accuracy. The result of research proved that CNN is the best model with 80% accuracy

Full Text Download

 

Paper 2: SLUM RESIDENTS AS SECONDARY CONTRIBUTORS IN MARKET ECONOMY: A REVIEW (7-11)

Abstract: The Slum areas are a globalized problem as several countries suffer from, especially the developing countries. According to UN estimations, almost one billion people reside in urban areas live under housing conditions that are characterized as slum areas or squatter settlements and substantially consider as poverty traps. Slum Residents plays an important role to the economies of developing countries. Hence it is necessary to balance the opportunities of secondary contributors such as labors, farmers and lower staff. The slum dwellers are always ignored. The barriers to slum residents constrain the economy of country. This review explore the massive contribution of slum labors in the human capital. Hence it is necessary to address the problems of secondary contributors to the growing economies of Pakistan.

Full Text Download

 

Paper 3: RAILWAY TRACK CONDITION MONITORING WITH ABA METHODOLOGY BY USING SMART IMU SENSORS (12-17)

Abstract: IMost of the railway tracks in Pakistan have surpassed their service lives and are unable to carry rail traffic load. To avoid any unwanted incidence, railway condition monitoring serves as a vital part in the fault detection. One of those crucial faults that is invisible to the image processing algorithms is railway track drainage issue, whose main causes can be residual of rainwater on the track surface, dew, moisture, snow, and seepage of water from adjacent areas. As a result, the level of moisture increases causing the partial erosion in the track’s embarkment, which consequently leads to the corrugation. This paper demonstrates how the early detection of railway track faults is possible via enhanced instrumentation based on axle box acceleration (ABA) and adequate postprocessing. Therefore, in this study an IoT enabled device is developed using Microcontroller Unit (Node MCU Esp-32), an accelerometer (ADXL345) and a 5V, 10000mAH battery which recognizes the track faults from the accelerometer, and a binary classifier shows the faulty and unfaulty conditions. The proposed system has been tested on Realtime environment whose response can be continuously monitored from anywhere.

Full Text Download

 

Paper 4: MALARIA CELLS PREDICTION OF COAL QUALITY PARAMETERS USING DIGITAL IMAGE PROCESSING (18-23)

Abstract: Quality assessment is of prime importance for the acceptance or rejection of coal in coal-fired power plants. Conventional coal quality assessment methods are time consuming due to the coal preparation and analysis which require multiple equipment. Based on color and texture, rapid coal assessment prediction tools have been developed to minimize the time, expenses and effort of coal assessment. In this research, multi-regression models were developed to predict Fixed Carbon (FC), Ash, and Gross Calorific Value (GCV) from a coal image through Digital Image Processing (DIP) and compared with the conventional coal quality assessment results. The simulation of DIP showed R2 values for ash vs features 72.6%, fixed carbon vs features is 70.5% and GCV vs features 64.5%. From the results, it can be concluded that the relation between FC and ash with coal image features is more justified than GCV. The R2 values for Ash and FC were found better and could be used to predict coal quality parameters such as FC and Ash content for a particular coal extracted from Lakhra Coal Mines. Meanwhile, the possibility of this multi-regression models would be validated for various coal samples of indigenous deposits.

Full Text Download