Work Efficiency Prediction Analysis and Optimal Path Finding Algorithm
A correct prediction of a contractor’s work can lead to many important things like keeping corruption in check, determining whether that particular is fit for the job or not. Frequently, it is brought that prediction is chaotic rather than random, which means it can be predicted by carefully analysing the history of respective contractor. Machine learning is an efficient way to represent such processes. It predicts a value close to the tangible value, thereby increasing the accuracy. The vital part of machine learning is the dataset used. The dataset should be as concrete as possible because a little change in the data can perpetuate massive changes in the outcome. It consists of variables like name of contactors, previous projects that he worked on, budget of that project, estimated lifespan of the project and actual life of the project. Due to the serious problem of bad road construction and maintenance, the real-time road situation and the possibility of road conditions in the next time period should be taken into accounts through the vehicle navigation system, in order to provide the optimal routing plans for vehicles in routing optimization. To solve the ignoring of real-time travel information and historical travel information in the existing navigation systems during routing optimization, this paper compares the existing navigation system to a prototype system which considers the conditions of the roadways as one of the factor in the recommendation and optimization system of the path finding and navigation process.
Published by: Rohit Pillai, Vipul Bhangale, Bishal Anand, Adarsh Kumar
Author: Rohit Pillai
Paper ID: V7I3-1766
Paper Status: published
Published: June 10, 2021
RICA: Real-Time Image Captioning Application
Automatically describing the content of an image is a fundamental problem in artificial intelligence that connects computer vision and natural language processing. Image caption generator is a task that involves computer vision and natural language processing concepts to recognize the context of an image and describe them in a natural language like English. The recent advances in Deep Learning-based Machine Translation and Computer Vision have led to excellent Image Captioning models using advanced techniques like Deep Reinforcement Learning. While these models are very accurate, these often rely on the use of expensive computation hardware making it difficult to apply these models in real-time scenarios, where their actual applications can be realized. In this paper, we carefully follow some of the core concepts of Image Captioning and its common approaches and present our simplistic encoder and decoder-based implementation with significant modifications and optimizations which enable us to run these models on low-end hardware of hand-held devices. We also compare our results evaluated using various metrics with state-of-the-art models and analyze why and where our model trained on the MSCOCO dataset lacks due to the trade-off between computation speed and quality. Using the state-of-the-art TensorFlow framework by Google, we also implement a first-of-its-kind Android application to demonstrate the real-time applicability and optimizations of our approach.
Published by: Suraj Dahake, Aditya Ohekar, Shubham Ilag, Aasim Shah
Author: Suraj Dahake
Paper ID: V7I3-1729
Paper Status: published
Published: June 10, 2021
Design of a Micro-controller based programmable voltage source for process automation
Automated control of processes are emerging in all spheres as it saves time and energy by providing a significant improvement in productivity and accuracy when compared to the manual methods of controlling processes which require constant human intervention. This paper presents a design of a voltage source that can be programmed using a micro-controller board to control the voltage that needs to be generated. The design is achieved by interfacing the electronic components required along with necessary software tools as a part of an embedded system consisting of the input/output peripherals and a microcontroller. There are two approaches of design, the first approach is based on a Digital to Analog converter and an Operational amplifier, the second approach is based on a MOSFET driver circuit. The circuits are constructed based on the design methodology which is developed with all the components chosen according to desired requirements. The circuits are simulated in Proteus in order to evaluate their performance and the readings are tabulated. The accuracy of both the approaches are computed and it is found that both the approaches have high levels of accuracy.
Published by: Rahul Desingh S., Sindhu Rajendran
Author: Rahul Desingh S.
Paper ID: V7I3-1774
Paper Status: published
Published: June 10, 2021
Self-wound analysis using Machine Learning and Image Processing
The significance of powerful surgical wound care can never be underestimated. Poorly managing surgical wounds may reason many critical complications. As a result, it increases. The necessity to broaden a patient-friendly self-care device which can assist both sufferers and clinical specialists to ensure the Nation of the surgical wounds without any unique medical equipment. On this paper, a surgical wound evaluation gadget for Self-care is proposed. The proposed machine is designed to allow patients seize surgical wound pictures of themselves with the aid of the usage of a cellular tool and add these pix for evaluation. Combining Image-processing and gadget-gaining knowledge of strategies, the proposed approach consists of four levels. First, photos are segmented into superpixels wherein each superpixel carries the pixels within the comparable shade distribution. 2nd, these superpixels corresponding to the pores and skin are recognized and the area of related skin Superpixels is derived. 1/3, surgical wounds can be extracted from this place primarily based on the statement of the texture distinction between skin and wounds. Ultimately, country and signs and symptoms of surgical Wound may be assessed. Full-size experimental effects are Conducted. With the proposed method, greater than 90% of country evaluation consequences are correct, and greater than ninety one% of symptom evaluation results consistent with the real analysis. Furthermore, case studies are furnished to show the benefit and trouble of this machine. Those results display that this device should perform
Published by: Varun Ganesh, U. Nomesh
Author: Varun Ganesh
Paper ID: V7I3-1767
Paper Status: published
Published: June 10, 2021
ERP system for large scale business vs small & mid-scale business
Enterprise resource planning (ERP) System is being used for the last 15 to 20 years and This research paper attempts to identify and explore the main issues affecting large scale businesses as well as small and medium scale businesses regarding the implementation of ERP System in enterprises. In this paper manufacturing companies are ranging in size from million dollars to over billion dollars as per their annual turnover. Basically, this research paper highlights the main issues where the different methods or solutions are to be applied according to the organization's scale. Except this also the benefits get differed by company size as small and medium companies gets some limitations compared with large companies. Large companies get financial improvement and economically better results whereas these smaller and mid-size organizations get better results in Production as well Logistics and manufacturing sectors.
Published by: Savita Virkar, Lavina Jadhav
Author: Savita Virkar
Paper ID: V7I3-1761
Paper Status: published
Published: June 10, 2021
Enriching Indoor and outdoor Fire Detection through CNN
Fire is a highly useful as well as a dangerous resource that has been utilized by humans since centuries. The only type of fire that is highly useful is the type of fire that is controlled and the energy generated can be used for different purposes. But not all fires are like that and some fires can be extremely devastating. These fires can become large and take down acres and acres of forests leading to extreme death and destruction. There have been recent and highly devastating fires that have rocked major rainforests and decimated a lot of wildlife close to extinction and endangerment. These fires can be stopped if detected when they are in their starting stages and lead to effective reduction in the destruction. There are several techniques such as sensors and other equipment that have been useful in the detection of the fire, but they have not been highly effective and efficient in the deployment. Therefore, an image processing based approach is defined in this research article to achieve effective realization of the fire detection. The proposed approach utilizes Convolutional Neural Networks along with Decision Tree to achieve the effective Fire detection. The experimental results confirm the accurate deployment of the fire detection mechanism.
Published by: Swapnali Kamble, Vaishnavi Sade, Rutuja Kamble, Sumedh Patil, Shubhangi Ingale, Shalaka Deore
Author: Swapnali Kamble
Paper ID: V7I3-1752
Paper Status: published
Published: June 10, 2021
