Two Stage Semantic Segmentation by SEEDS and Fork Net

Mukherjee, Aritra; Jana, Pulak Kumar; Chakraborty, Sourav; Saha, Sanjoy Kumar

doi:10.1109/calcon49167.2020.9106468

Cited by 7 publications

(5 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Along with this, Mask R-CNN has an additional parallel pipeline that generates object masks. These masks are similar to performing semantic segmentation [31] on the whole image to generate pixel-wise classification. RoIPool serves extraction of feature map from each region of interest.…”

Section: Region Proposal-based Two-stage Architectures For 2d Object ...mentioning

confidence: 99%

Recent Trends in 2D Object Detection and Applications in Video Event Recognition

Jana¹,

Mohanta²

2022

Preprint

Self Cite

View full text Add to dashboard Cite

Object detection serves as a significant step in improving performance of complex downstream computer vision tasks. It has been extensively studied for many years now and current state-of-the-art 2D object detection techniques proffer superlative results even in complex images. In this chapter, we discuss the geometry-based pioneering works in object detection, followed by the recent breakthroughs that employ deep learning. Some of these use a monolithic architecture that takes a RGB image as input and passes it to a feed-forward ConvNet or vision Transformer. These methods, thereby predict class-probability and bounding-box coordinates, all in a single unified pipeline. Two-stage architectures on the other hand, first generate region proposals and then feed it to a CNN to extract features and predict object category and boundingbox. We also elaborate upon the applications of object detection in video event recognition, to achieve better fine-grained video classification performance. Further, we highlight recent datasets for 2D object detection both in images and videos, and present a comparative performance summary of various state-of-the-art object detection techniques.

show abstract

Section: Region Proposal-based Two-stage Architectures For 2d Object ...mentioning

confidence: 99%

Recent Trends in 2D Object Detection and Applications in Video Event Recognition

Jana¹,

Mohanta²

2022

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…Some of the initial CNN networks, intuitively exploited the spatial features that the videos contained, as the models were fed with successive frames of the video. Alternatively, this was also effectuated by performing image classification and/or semantic segmentation Mukherjee et al (2020) on the individual frames. Such spatial techniques can correlate a scene's subject to the background, thereby realizing the scene content; as for instance, the common appearances of a soccer ball in a stadium or gallery, and the fact that a football match could be prevalent in the event.…”

Section: Recent Approaches Using Deep Neural Architecturesmentioning

confidence: 99%

Event and Activity Recognition in Video Surveillance for Cyber-Physical Systems

Bhaumik¹,

Jana

Mohanta

2021

Advances in Science, Technology &Amp; Innovation

Self Cite

View full text Add to dashboard Cite

In this chapter, we aim to aid the development of Cyber-Physical Systems (CPS) in automated understanding of events and activities in various applications of videosurveillance. These events are mostly captured by drones, CCTVs or novice and unskilled individuals on low-end devices. Being unconstrained in nature, these videos are immensely challenging due to a number of quality factors. We present an extensive account of the various approaches taken to solve the problem over the years. This ranges from methods as early as Structure from Motion (SFM) based approaches to recent solution frameworks involving deep neural networks. We show that the long-term motion patterns alone play a pivotal role in the task of recognizing an event. Consequently each video is significantly represented by a fixed number of key-frames using a graph-based approach. Only the temporal features are exploited using a hybrid Convolutional Neural Network (CNN) + Recurrent Neural Network (RNN) architecture. The results we obtain are encouraging as they outperform standard temporal CNNs and are at par with those using spatial information along with motion cues. Further exploring multistream models, we conceive a multi-tier fusion strategy for the spatial and temporal wings of a network. A consolidated representation of the respective individual prediction vectors on video and frame levels is obtained

show abstract

“…Mutherjee at el. [11] worked on semantic segmentation on images to do an effective annotation at pixel-level. Thus a roadmap with erected buildings and monuments of a city-scape or village-scape can be semantically segmented to a map, and eventually to a graph in Euclidean space.…”

Section: Related Workmentioning

confidence: 99%

“…From that position, the drone's camera can capture high-resolution RGB images that represents a bird's eyeview of the extended region surrounding the drone. Fast semantic segmentation on RGB images through fully-convolutional network [16] or superpixel-based classification [11] can label each pixel to their corresponding class as building, tree, road, tower, etc. Further, panoptic segmentation [17] separates different instances of the same object (e.g.…”

Section: Perception and Real-time Formation Of Mapsmentioning

confidence: 99%

AutoDrone: Shortest Optimized Obstacle-Free Path Planning for Autonomous Drones

Jana¹,

Jana²

2021

Preprint

Self Cite

View full text Add to dashboard Cite

With technological advancement, drone has emerged as unmanned aerial vehicle that can be controlled by humans to fly or reach a destination. This may be autonomous as well, where the drone itself is intelligent enough to find a shortest obstacle-free path to reach the destination from a designated source. Be it a planned smart city or even a wreckage site affected by natural calamity, we may imagine the buildings, any surface-erected structure or other blockage as obstacles for the drone to fly in a straight line-of-sight path. To address such path-planning of drones, the bird's eye-view of the whole landscape is first transformed to a graph of grid-cells, where some are occupied to indicate the obstacles and some are free to indicate the free path. We propose a method to find out the shortest obstacle-free path in the coordinate system guided by GPS. The autonomous drone (AutoDrone) will thus be able to move from one place to another along the shortest path, without colliding into hindrances, while traveling in a two-dimensional space. Heuristics to extend this to long journeys and 3D space are also elaborated. Our approach can be especially beneficial in rescue operations and fast delivery or pick-up in an energy-efficient way, where our algorithm will help in finding out the shortest path and angle along which it should fly. Experiments are done on different scenarios of map layouts and obstacle positions to understand the shortest feasible route, computed by the autonomous drone.

show abstract

Two Stage Semantic Segmentation by SEEDS and Fork Net

Cited by 7 publications

References 25 publications

Recent Trends in 2D Object Detection and Applications in Video Event Recognition

Recent Trends in 2D Object Detection and Applications in Video Event Recognition

Event and Activity Recognition in Video Surveillance for Cyber-Physical Systems

AutoDrone: Shortest Optimized Obstacle-Free Path Planning for Autonomous Drones

Contact Info

Product

Resources

About