An Efficient Tamper Detection System for End to End Video Communication Using Wavelet

2017 ◽  
Vol 14 (3) ◽  
pp. 1493-1498
Author(s):  
A. Velmurugan ◽  
E. Logashanmugam
Keyword(s):  
Detection System ◽  
Tamper Detection ◽  
Video Communication ◽  
End To End

scholarly journals EHSOD: CAM-Guided End-to-End Hybrid-Supervised Object Detection with Cascade Refinement

2020 ◽  
Vol 34 (07) ◽  
pp. 10778-10785
Author(s):  
Linpu Fang ◽  
Hang Xu ◽  
Zhili Liu ◽  
Sarah Parisot ◽  
Zhenguo Li
Keyword(s):  
Object Detection ◽  
State Of The Art ◽  
Detection System ◽  
Parameter Tuning ◽  
Heat Map ◽  
Level Data ◽  
End To End ◽  
Weakly Supervised ◽  
Yield State ◽  
Activation Heat

Object detectors trained on fully-annotated data currently yield state of the art performance but require expensive manual annotations. On the other hand, weakly-supervised detectors have much lower performance and cannot be used reliably in a realistic setting. In this paper, we study the hybrid-supervised object detection problem, aiming to train a high quality detector with only a limited amount of fully-annotated data and fully exploiting cheap data with image-level labels. State of the art methods typically propose an iterative approach, alternating between generating pseudo-labels and updating a detector. This paradigm requires careful manual hyper-parameter tuning for mining good pseudo labels at each round and is quite time-consuming. To address these issues, we present EHSOD, an end-to-end hybrid-supervised object detection system which can be trained in one shot on both fully and weakly-annotated data. Specifically, based on a two-stage detector, we proposed two modules to fully utilize the information from both kinds of labels: 1) CAM-RPN module aims at finding foreground proposals guided by a class activation heat-map; 2) hybrid-supervised cascade module further refines the bounding-box position and classification with the help of an auxiliary head compatible with image-level data. Extensive experiments demonstrate the effectiveness of the proposed method and it achieves comparable results on multiple object detection benchmarks with only 30% fully-annotated data, e.g. 37.5% mAP on COCO. We will release the code and the trained models.

scholarly journals Nanowire Ring Embedded in a Flexible Substrate for Local Strain Detection

Materials ◽  
2020 ◽  
Vol 13 (2) ◽  
pp. 347
Author(s):  
Shengkun Li ◽  
Yue Qin ◽  
Xin Li ◽  
Yuejin Zhao
Keyword(s):  
Detection System ◽  
Structural Parameters ◽  
Flexible Substrate ◽  
Ring Cavity ◽  
Local Strain ◽  
Strain Sensors ◽  
Resonant Wavelength ◽  
Gauge Factor ◽  
Light Sources ◽  
End To End

Optical sensing has attracted more and more attention in recent years with the advance in planar waveguide fabrication processes. The photon, as a carrier of information in sensing areas, could have a better performance than electrons. We propose a novel end-to-end ring cavity to fabricate sensitive units of a strain sensor. We then propose a method of combining a flexible substrate with an end-to-end semiconductor nanowire ring cavity to fabricate novel strain sensors. We used a tuning resonant wavelength detected by a homebuilt excitation and detection system to measure applied strain. The resonant wavelength of the strain gauge was red-shift and linear tuned with increasing strain. The gauge factor was about 50, calculated through experiments and theory, and Q was 1938, with structural parameters L = 70 µm and d = 1 µm. The high sensitivity makes it possible to measure micro deformation more accurately. End-to-end coupling active nanowire waveguides eliminate the shortcomings of side by side coupling structures, which have the phasing shift with no minor optical density loss. This resonator in flexible substrates could be used not only as on-chip strain sensors for micro or nano deformation detecting but also as tunable light sources for photonic integrated circuits.

scholarly journals An end-to-end approach for optimal mode selection in Internet video communication: theory and application

2000 ◽  
Vol 18 (6) ◽  
pp. 977-995 ◽  
Author(s):  
D. Wu ◽  
Y.T. Hou ◽  
B. Li ◽  
Wenwu Zhu ◽  
Ya-Qin Zhang ◽  
...  
Keyword(s):  
Communication Theory ◽  
Mode Selection ◽  
Video Communication ◽  
Internet Video ◽  
Optimal Mode ◽  
End To End
Sign in / Sign up

Export Citation Format

Share Document