AMMDAS: Multi-Modular Generative Masks Processing Architecture With Adaptive Wide Field-of-View Modeling Strategy

Venkata Subbaiah Desanamukula,Preethi Padala,Pushkal Padala,Prasad Reddy Pvgd,Premith Kumar Chilukuri

doi:10.1109/access.2020.3033537

Venkata Subbaiah Desanamukula, Preethi Padala + Show 3 more

Open Access

https://doi.org/10.1109/access.2020.3033537

Copy DOI

Abstract

The usage of transportation systems is inevitable; any assistance module which can catalyze the flow involved in transportation systems, parallelly improving the reliability of processes involved is a boon for day-to-day human lives. This paper introduces a novel, cost-effective, and highly responsive Post-active Driving Assistance System, which is "Adaptive-Mask-Modelling Driving Assistance System" with intuitive wide field-of-view modeling architecture. The proposed system is a vision-based approach, which processes a panoramic-front view (stitched from temporal synchronous left, right stereo camera feed) & simple monocular-rear view to generate robust & reliable proximity triggers along with co-relative navigation suggestions. The proposed system generates robust objects, adaptive field-of-view masks using FRCNN+Resnet-101_FPN, DSED neural-networks, and are later processed and mutually analyzed at respective stages to trigger proximity alerts and frame reliable navigation suggestions. The proposed DSED network is an Encoder-Decoder-Convolutional-Neural-Network to estimate lane-offset parameters which are responsible for adaptive modeling of field-of-view range (157 o -210 o ) during live inference. Proposed stages, deep-neural-networks, and implemented algorithms, modules are state-of-the-art and achieved outstanding performance with minimal loss(L{p, t}, L δ , L Total ) values during benchmarking analysis on our custombuilt, KITTI, MS-COCO, Pascal-VOC, Make-3D datasets. The proposed assistance-system is tested on our custom-built, multiple public datasets to generalize its reliability and robustness under multiple wild conditions, input traffic scenarios & locations.

Highlights

Transportation plays an indispensable role in individual and social welfare, the economy, and quality of life
To tackle the above case we introduce an extra submodule(Stage-5), where the input 1650 panoramic image is directly consumed in order generate a dense depth map [17], [14], [11]; and robust navigation suggestions are framed by processing these dense depth maps
EXPERIMENTS, RESULTS & DISCUSSION We have tested and experimented the proposed system in live traffic recordings, where an initial setup of 2 smartphones or web-cams are mounted to a stand separated by 120cm distance, the left and right stereo feeds are processed by stages 1-5 to output corresponding dynamic proximity alerts along with relative adaptive navigation suggestions (Tables 4 and 5)

Summary

Introduction

Transportation plays an indispensable role in individual and social welfare, the economy, and quality of life. Recent studies from the World Health Organization indicate that 1.25 million deaths occur every year due to road traffic accidents. Recent studies from the World Health Organization indicate that 1.25 million deaths occur every year due to road traffic accidents1 Such accidents resulted in a global cost of ~US$518 billion per year, which results in a decline of ~1-2% gross domestic product from all of the countries in the world2 [1]. The tracking and recognition of traffic participants (pedestrians, cars, bicyclists [3]–[7], etc) which are in proximity plays a crucial role in the safe maneuvering of self-governed autonomous [8] vehicles

Methods

Results

Conclusion