Focus: Querying Large Video Datasets with Low Latency and Low Cost

Abstract

Large volumes of videos are continuously recorded from cameras deployed for traffic control and surveillance with the goal of answering "after the fact" queries: identify video frames with objects of certain classes (cars, bags) from many days of recorded video. While advancements in convolutional neural networks (CNNs) have enabled answering such queries with high accuracy, they are too expensive and slow. We build Focus, a system for low-latency and low-cost querying on large video datasets. Focus uses cheap ingestion techniques to index the videos by the objects occurring in them. At ingest-time, it uses compression and video-specific specialization of CNNs. Focus handles the lower accuracy of the cheap CNNs by judiciously leveraging expensive CNNs at query-time. To reduce query time latency, we cluster similar objects and hence avoid redundant processing. Using experiments on video streams from traffic, surveillance and news channels, we see that Focus uses 58X fewer GPU cycles than running expensive ingest processors and is 37X faster than processing all the video at query time.

References in corpus (9)

• Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
• Distilling the Knowledge in a Neural Network
• FitNets: Hints for Thin Deep Nets
• Exploiting Linear Structure Within Convolutional Networks for Efficient Evaluation
• CNN Features off-the-shelf: an Astounding Baseline for Recognition
• Speeding up Convolutional Neural Networks with Low Rank Expansions
• YOLO9000: Better, Faster, Stronger
• Aggregating Deep Convolutional Features for Image Retrieval
• Learning Complexity-Aware Cascades for Deep Pedestrian Detection

Cited by in corpus (14)

• SPINN: Synergistic Progressive Inference of Neural Networks over Device and Cloud
• MobiSR: Efficient On-Device Super-Resolution through Heterogeneous Mobile Processors
• The Internet of Federated Things (IoFT): A Vision for the Future and In-depth Survey of Data-driven Approaches for Federated Learning
• VidCEP: Complex Event Processing Framework to Detect Spatiotemporal Patterns in Video Streams
• VID-WIN: Fast Video Event Matching with Query-Aware Windowing at the Edge for the Internet of Multimedia Things
• ILCAS: Imitation Learning-Based Configuration-Adaptive Streaming for Live Video Analytics with Cross-Camera Collaboration
• Minimum Viable Device Drivers for ARM TrustZone
• Extract-Transform-Load for Video Streams
• ExSample: Efficient Searches on Video Repositories through Adaptive Sampling
• ApproxNet: Content and Contention-Aware Video Analytics System for Embedded Clients
• Jointly Optimizing Preprocessing and Inference for DNN-based Visual Analytics
• ReXCam: Resource-Efficient, Cross-Camera Video Analytics at Scale
• Scaling Video Analytics Systems to Large Camera Deployments
• Pool of Experts: Realtime Querying Specialized Knowledge in Massive Neural Networks