| 1. |
- Ahmed, Soban, et al.
(författare)
-
Development and Optimization of Deep Learning Models for Weapon Detection in Surveillance Videos
- 2022
-
Ingår i: Applied Sciences. - : MDPI. - 2076-3417. ; 12:12
-
Tidskriftsartikel (refereegranskat)abstract
- Featured Application This work has applied computer vision and deep learning technology to develop a real-time weapon detector system and tested it on different computing devices for large-scale deployment. Weapon detection in CCTV camera surveillance videos is a challenging task and its importance is increasing because of the availability and easy access of weapons in the market. This becomes a big problem when weapons go into the wrong hands and are often misused. Advances in computer vision and object detection are enabling us to detect weapons in live videos without human intervention and, in turn, intelligent decisions can be made to protect people from dangerous situations. In this article, we have developed and presented an improved real-time weapon detection system that shows a higher mean average precision (mAP) score and better inference time performance compared to the previously proposed approaches in the literature. Using a custom weapons dataset, we implemented a state-of-the-art Scaled-YOLOv4 model that resulted in a 92.1 mAP score and frames per second (FPS) of 85.7 on a high-performance GPU (RTX 2080TI). Furthermore, to achieve the benefits of lower latency, higher throughput, and improved privacy, we optimized our model for implementation on a popular edge-computing device (Jetson Nano GPU) with the TensorRT network optimizer. We have also performed a comparative analysis of the previous weapon detector with our presented model using different CPU and GPU machines that fulfill the purpose of this work, making the selection of model and computing device easier for the users for deployment in a real-time scenario. The analysis shows that our presented models result in improved mAP scores on high-performance GPUs (such as RTX 2080TI), as well as on low-cost edge computing GPUs (such as Jetson Nano) for weapon detection in live CCTV camera surveillance videos.
|
|
| 2. |
- Dost, Shahi, et al.
(författare)
-
Reduced reference image and video quality assessments : review of methods
- 2022
-
Ingår i: EURASIP Journal on Image and Video Processing. - : Springer. - 1687-5176 .- 1687-5281. ; 2022:1
-
Forskningsöversikt (refereegranskat)abstract
- With the growing demand for image and video-based applications, the requirements of consistent quality assessment metrics of image and video have increased. Different approaches have been proposed in the literature to estimate the perceptual quality of images and videos. These approaches can be divided into three main categories; full reference (FR), reduced reference (RR) and no-reference (NR). In RR methods, instead of providing the original image or video as a reference, we need to provide certain features (i.e., texture, edges, etc.) of the original image or video for quality assessment. During the last decade, RR-based quality assessment has been a popular research area for a variety of applications such as social media, online games, and video streaming. In this paper, we present review and classification of the latest research work on RR-based image and video quality assessment. We have also summarized different databases used in the field of 2D and 3D image and video quality assessment. This paper would be helpful for specialists and researchers to stay well-informed about recent progress of RR-based image and video quality assessment. The review and classification presented in this paper will also be useful to gain understanding of multimedia quality assessment and state-of-the-art approaches used for the analysis. In addition, it will help the reader select appropriate quality assessment methods and parameters for their respective applications.
|
|
| 3. |
- Khan, Muhammad Gufran, et al.
(författare)
-
A Doublet-Shift Transmitted Reference Scheme for Ultra-Wideband Communication Systems
- 2007
-
Konferensbidrag (refereegranskat)abstract
- The transmitted reference (TR) scheme has gained attention for ultra-wideband (UWB) communications as it bypasses the complex task of channel estimation and provides a simple receiver structure. However, the low-complexity detection in this scheme is achieved at the expense of a loss in performance. To this end, this paper proposes a doublet-shift TR (DsTR) signaling and detection scheme to improve the performance of conventional TR impulse radio UWB system. The performance of the proposed scheme is compared with conventional TR signaling scheme in terms of uncoded bit error rate (BER). The simulation results validate that the doublet-shift TR scheme achieves about 1.5-2 dB better performance than the conventional TR scheme. The DsTR scheme does not require extra energy and still has the low-complexity.
|
|
| 4. |
- Khan, Muhammad Gufran, et al.
(författare)
-
An energy-efficient signaling and detection scheme for transmitted reference UWB systems
- 2008
-
Konferensbidrag (refereegranskat)abstract
- The transmitted reference (TR) schemes do not require the complex task of channel estimation and provide a simple receiver structure. However, the low-complexity detection in conventional TR scheme is achieved at the expense of 50% energy/rate loss. In this paper, an energy-efficient TR signaling and detection scheme is presented with the aim to improve the performance of conventional TR impulse radio UWB system. The proposed scheme is a combination of the conventional TR scheme and the differential TR scheme. The simulation results validate that the proposed TR scheme utilizes 1.76 dB less energy per bit and recovers 25% rate loss of conventional TR scheme, while maintaining the same BER performance. Further, the proposed ATR scheme improves the BER performance by about 2 dB using simple averaging of the reference pulses. ©2008 IEEE.
|
|
| 5. |
- Khan, Muhammad Gufran, et al.
(författare)
-
Detection of Impulse Radio Ultra-Wideband Signals using Recursive Transmitted Reference Receivers
- 2007
-
Konferensbidrag (refereegranskat)abstract
- For the detection of impulse radio (IR) UWB signals, complex RAKE receiver or low-complexity transmitted reference (TR) autocorrelation receivers can be used. However, there is a performance loss associated with the low-complexity TR scheme. This paper proposes recursive structures of the conventional TR and averaged TR schemes to improve the detection performance of IR-UWB signals. The performance of proposed schemes is evaluated over the standard IEEE 802.15.4a multipath channels. The performance is compared with conventional TR receivers in terms of uncoded bit-error-rate (BER), assuming that the channel is quasi-static. For averaged and recursive averaged TR schemes, the TR sequence is also slightly modified. The simulation results validate that the proposed schemes has better performance by about 2 dB than the conventional TR and averaged TR receivers.
|
|
| 6. |
- Khan, Muhammad Gufran, et al.
(författare)
-
Energy efficient binary PPM/code-multiplexed transmitted-reference multi-user UWB system
- 2011
-
Konferensbidrag (refereegranskat)abstract
- This paper presents a combined binary pulse position modulation (BPPM)/code-multiplexed transmitted-reference (CM-TR) ultra wideband (UWB) system. The low complexity CM-TR UWB is not energy efficient as one half of the energy is used to transmit reference pulses and, even at low-to-medium data rates, it suffers from strong multi-user interference (MUI) and inter-frame interference (IFI). Keeping the information rate constant, the combined BPPM/CM-TR UWB system utilizes 3 dB less energy per bit as compared to the CM-TR system and achieves relatively longer silent periods between the pulses, which not only mitigate the IFI but also decrease the MUI due to less collisions with other asynchronous users. In the high SNR region, where the effect of IFI and MUI severely degrades the bit-error-rate (BER) performance of the multiuser CM-TR system, simulation results verify that the multiuser BPPM/CM-TR outperforms the multiuser CM-TR system.
|
|
| 7. |
- Khan, Muhammad Gufran, et al.
(författare)
-
Measurements and Analysis of UWB Channels in Industrial Environments
- 2005
-
Konferensbidrag (refereegranskat)abstract
- Communication in industrial environments is an important scenario for ultra wideband (UWB) systems. The characteristics of this environment differ considerably from office environments, where most of the channel measurements reported in the literature are performed. In this paper, we present the results and analysis of UWB channel measurements in an industrial environment. We investigate both peer-to-peer non-line-of-sight (P-P NLOS) and base station NLOS (BS NLOS) scenarios. The measurements are performed in the frequency range 3.1 to 8 GHz, at distances 2-16 m, using a vector network analyzer in conjunction with virtual antenna arrays. From these measurements, we extract channel model parameters such as power delay profiles, pathloss exponents, rms delay spread and the percentage of energy captured by RAKE receivers with a finite number of fingers. In addition, Saleh- Valenzuela parameters such as ray and cluster power decay constants and distribution of cluster interarrival times are extracted.
|
|
| 8. |
- Khan, Muhammad Gufran, et al.
(författare)
-
Non-coherent detection of impulse radio UWB signals based on fourth order statistics
- 2009
-
Konferensbidrag (refereegranskat)abstract
- Non-coherent receivers are an attractive solution for impulse radio (IR) ultra wideband (UWB) systems due to their low-complexity implementation and low power consumption. However, low-complexity detection is achieved at the expense of some loss in performance. In this paper, a non-coherent detection scheme based on fourth order statistics, i.e., kurtosis, is proposed for low data rate IR-UWB systems. The performance of a conventional energy detector (ED) and the proposed kurtosis detector (KD) is compared in terms of bit-error-rate (BER). The simulation results validate that the KD receiver outperforms a conventional ED receiver on both additive white Gaussian noise (AWGN) and IEEE 802.15.4a multipath channels.
|
|
| 9. |
- Khan, Muhammad Gufran, et al.
(författare)
-
Non-coherent detection of impulse radio UWB signals based on fourth order statistics
- 2011
-
Ingår i: Wireless Personal Communications. - : Springer Science and Business Media LLC. - 1572-834X .- 0929-6212.
-
Tidskriftsartikel (refereegranskat)abstract
- Low-complex and low power non-coherent energy detector (ED) is interesting for low data rate impulse radio (IR) ultra wideband (UWB) systems but, compared to coherent receivers, it suffers from a loss in performance due to low signal-to-noise ratio (SNR) at the detector. In addition, the performance of an ED strongly depends on the integration interval (window size) of the integrator and the window position. A non-coherent kurtosis detector (KD) and a fourth-order detector (FD), which can discriminate between Gaussian noise signals and non-Gaussian IR-UWB signals by directly estimating the fourth-order moment of the received signal, are presented. The performance of the detectors is evaluated using real channels measured in a corridor, an office and a laboratory environment. The results show that bit-error-rate (BER) performance of the proposed KD receiver is better than the ED receiver only under certain conditions, while the FD receiver is slightly better than the ED in low SNR region and its performance improves as the SNR increases. In addition, the performance of the FD receiver is less sensitive to overestimation of the integration interval making it relatively robust to variations of the channel delay spread. Finally, a criteria for the selection of integration time of the proposed detector is suggested.
|
|
| 10. |
- Khan, Muhammad Gufran, et al.
(författare)
-
Non-Coherent Fourth-Order Detector for Impulse Radio Ultra Wideband Systems : Empirical Evaluation Using Channel Measurements
- 2013
-
Ingår i: Wireless personal communications. - : Springer. - 0929-6212 .- 1572-834X. ; 68:1, s. 27-46
-
Tidskriftsartikel (refereegranskat)abstract
- Low-complex and low-power non-coherent energy detectors (EDs) are interesting for low data rate impulse radio (IR) ultra wideband (UWB) systems, but suffer from a loss in performance compared to coherent receivers. The performance of an ED also strongly depends on the integration interval (window size) of the integrator and the window position. This paper presents a non-coherent fourth-order detector (FD) which can discriminate between Gaussian noise signals and non-Gaussian IR-UWB signals by directly estimating the fourth-order moment of the received signal. The performance of the detectors is evaluated using realistic channels measured in a corridor, an office and a laboratory environment. The results show that bit-error-rate (BER) performance of the proposed FD receiver is slightly better than the ED in low signal-to-noise ratio (SNR) region and its performance improves as the SNR increases. In addition, BER of the FD receiver is less sensitive to overestimation of the integration interval making it relatively robust to variations of the channel delay spread. Finally, a criteria for the selection of integration time of the proposed detector is suggested.
|
|
