State Summarization of Video Streams for Spatiotemporal Query Matching in Complex Event Processing

Yadav, Piyush; Das, Dibya Prakash; Curry, Edward

doi:10.1109/icmla.2019.00022

Cited by 5 publications

(3 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The framework can be deployed over the cloud and can be exposed as an API to provide services over OSM. More complex and potential traffic services like Vehicle Overtake and Lane Change [30] can be queried by creating operators and deploying them to the system. Our method relies on existing cameras installed along road networks and most of the limitations arise from camera deployment.…”

Section: Discussion Of Limitationsmentioning

confidence: 99%

Traffic Prediction Framework for OpenStreetMap using Deep Learning based Complex Event Processing and Open Traffic Cameras

Yadav,

Sarkar,

Salwala

et al. 2020

Preprint

Self Cite

View full text Add to dashboard Cite

Displaying near-real-time traffic information is a useful feature of digital navigation maps. However, most commercial providers rely on privacy-compromising measures such as deriving location information from cellphones to estimate traffic. The lack of an open-source traffic estimation method using open data platforms is a bottleneck for building sophisticated navigation services on top of OpenStreetMap (OSM). We propose a deep learning-based Complex Event Processing (CEP) method that relies on publicly available video camera streams for traffic estimation. The proposed framework performs near-real-time object detection and objects property extraction across camera clusters in parallel to derive multiple measures related to traffic with the results visualized on OpenStreetMap. The estimation of object properties (e.g. vehicle speed, count, direction) provides multidimensional data that can be leveraged to create metrics and visualization for congestion beyond commonly used density-based measures. Our approach couples both flow and count measures during interpolation by considering each vehicle as a sample point and their speed as weight. We demonstrate multidimensional traffic metrics (e.g. flow rate, congestion estimation) over OSM by processing 22 traffic cameras from London streets. The system achieves a near-real-time performance of 1.42 seconds median latency and an average F-score of 0.80.

show abstract

Section: Discussion Of Limitationsmentioning

confidence: 99%

Traffic Prediction Framework for OpenStreetMap using Deep Learning based Complex Event Processing and Open Traffic Cameras

Yadav,

Sarkar,

Salwala

et al. 2020

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…It facilitates the execution of pre-defined visual queries that make decisions without any human intervention. The applicability of GNOSIS is manifold, where multiple urban analytics-based ML/DNN models can be chained together as distributed services to detect complex spatio-temporal [12] urban event analytics patterns. The work demonstrates three tactile detection queries that set an initial pathway towards accessibility and assistive applications in the urban analytics domain.…”

Section: Introductionmentioning

confidence: 99%

UrbanAccess

Salwala

Yadav

Munirathnam

et al. 2021

Proceedings of the 1st International Workshop on Multimedia Computing for Urban Data

Self Cite

View full text Add to dashboard Cite

The smart city concept has now become one of the key enablers in urban city management. The adoption and permeation of ICT and AI-driven techniques have enabled the authorities to resolve poor urban planning issues with improved delivery of citizen services. Major urban problem is addressing the accessibility issue across cities road crossing and facilitating visually impaired people via well-defined infrastructure. The research presented in this paper emphasized urban analytics that studies the road crossings and challenges one faces when accessing the footpaths of a city using the Tactile surfaces. This work demonstrates a distributed event analytics platform-GNOSIS to detect complex accessibility event patterns. GNOSIS ingest video data streams from cities infrastructure such as CCTV and detect tactile surface event patterns using an ensemble of deep learning models using a declarative query language. The work analyzes mainly three types of tactile surface -Blister, Cycleway and Directional, collected from different cities in Ireland using crowd-sourcing techniques. GNOSIS makes decisions in real-time based on the type of tactile surface, colour and the making pattern.

show abstract

“…In the past few years, CNNs have been used as the principal approach to solving a variety of significant problems in cyber security [27], smart wearables [26], localization [29,30], resilient communication [17], consumer electronics [19], computer vision [28,31], etc. Although CNNs are renowned for their excellent performance, they are also computationally intensive, consuming hundreds of MB of memory and MFLOPS of computing power [20,32].…”

Section: Introductionmentioning

confidence: 99%

Rce-Nn

Sudharsan

Breslin

Ali³

2020

Proceedings of the 10th International Conference on the Internet of Things

View full text Add to dashboard Cite

Microcontroller Units (MCUs) in edge devices are resource constrained due to their limited memory footprint, fewer computation cores, and low clock speeds. These limitations constrain one from deploying and executing machine learning models on MCUs. To fit, deploy and execute Convolutional Neural Networks (CNNs) for any IoT use-case on small MCUs, a complete design flow is required. Resource Constrained Edge -Neural Networks (RCE-NN) is the name given to our proposed design flow, with a five-stage pipeline that developers can follow for executing CNNs on MCUs. In this pipeline, the initial model architecture and training stage consists of four well-defined tasks on model size, workload, operations and quantization awareness, which maps the desired CNN as captured in an executable specification to a resource-constrained MCU's specification. The next quantization and conversion stage reduces model size, saves memory, and simplifies calculations without much impact on the accuracy. In the third stage, the quantized version of the model is translated into a c-byte array since the MCUs lack native file-system support. The translated c-byte array is fused with the main program of an IoT use-case and binaries are built using techniques from the fourth stage. Finally, the method presented in the last deployment stage is used to flash the built binaries onto MCUs, as this method allows the memory of the MCU to be fully utilized by the CNN and its operations. We evaluated RCE-NN using eight popular MCU boards. The results show that, when users realize all five pipeline stages, they can fit, deploy and execute multiple CNNs across multiple open-source MCU boards. The RCE-NN pipeline components quantize and compress the CNNs to 1/10 ℎ of their original size, enabling the CNNs to fit on MCUs with no or minimal loss in performance, both after quantization and compression, and also during runtime.

show abstract

State Summarization of Video Streams for Spatiotemporal Query Matching in Complex Event Processing

Cited by 5 publications

References 20 publications

Traffic Prediction Framework for OpenStreetMap using Deep Learning based Complex Event Processing and Open Traffic Cameras

Traffic Prediction Framework for OpenStreetMap using Deep Learning based Complex Event Processing and Open Traffic Cameras

UrbanAccess

Rce-Nn

Contact Info

Product

Resources

About