The increased use of modern printing and scanning technologies has led to a significant rise in counterfeit currency production, posing a serious threat to global economies. To tackle this growing issue, our project, ...
详细信息
ISBN:
(纸本)9798350370249
The increased use of modern printing and scanning technologies has led to a significant rise in counterfeit currency production, posing a serious threat to global economies. To tackle this growing issue, our project, titled "Fake currency detection using Convolutional Neural Networks and imageprocessing," introduces an innovative solution that utilizes artificial intelligence (AI) and machine learning for efficient counterfeit detection. Financial institutions, banks, and businesses are facing heightened vulnerability to counterfeit currency, resulting in considerable financial losses and a decrease in the value of genuine money. Current currency detection systems often rely on time-consuming traditional methods and manual inspection, which are prone to human error. Even the counterfeit detection machines in use have limitations when it comes to identifying sophisticated counterfeit notes. Our project addresses these challenges by proposing an advanced system that integrates convolutional neural networks (CNNs) and imageprocessing techniques. Given the advancements in printing and scanning technologies, counterfeiting has evolved into a more sophisticated and widespread problem. Traditional currency detection methods, rooted in hardware and imageprocessing, have proven to be inefficient and time-consuming. Hence, there is a critical need for a more robust and rapid solution to detect counterfeit currency. Our proposed approach employs a transfer-learned CNN, a deep learning model trained on a dataset comprising real and fake currency images. The CNN learns the intricate features of both genuine and counterfeit banknotes, allowing it to accurately identify fake currency in real-time. The transfer learning process enables the CNN to leverage knowledge gained from a diverse dataset, improving its ability to recognize subtle patterns associated with counterfeit notes. The primary components of our project include a diverse dataset with images of real and fake currenc
Many studies have been conducted on the 3D reconstruction of subjects using multiple fixed cameras. Accepting the trade-off between the number of cameras and reconstruction quality, our studio is designed to capture h...
详细信息
Many studies have been conducted on the 3D reconstruction of subjects using multiple fixed cameras. Accepting the trade-off between the number of cameras and reconstruction quality, our studio is designed to capture high-quality models of one or two subjects for TV program use. Several cameras are mounted on a hemispherical dome with the stage in the center and a cloth cover on the frame for chroma-keying. The optimal camera numbers and placements for reconstruction were determined by simulation, and the 3D reconstruction was performed as a point cloud by a combination of visual hull and stereo matching. The quality was still not high enough, however, so we also added a surface light field to the point cloud to obtain the weighted average of rays from camera images close to the viewpoint. In the final stage, the images were then combined to the video, and errors generated during the reconstruction were compensated for by use of a deep neural network (DNN) for video translation. An offline processing studio has been built as a preliminary step towards real-timeprocessing, and the reconstructed 3D images have been evaluated subjectively for a number of subjects. These studies confirm the effectiveness of this studio design.
The proceedings contain 20 papers. The topics discussed include: edge deployed satellite image classification with TinEViT a X-Cube-AI compatible efficient vision transformer;a design space exploration framework for d...
ISBN:
(纸本)9781510673861
The proceedings contain 20 papers. The topics discussed include: edge deployed satellite image classification with TinEViT a X-Cube-AI compatible efficient vision transformer;a design space exploration framework for deployment of resource-constrained deep neural networks;IoT-enabled unmanned traffic management system with dynamic vision-based drone detection for sense and avoid coordination;exploring action recognition in endoscopy video datasets;integrating image-based LLMs on edge-devices for underwater robotics;age-based clustering of seagrass blades using AI models;layered convolutional neural networks for multi-class image classification;eyeball tracking in closed eyes from shadows;CAEN: efficient adversarial robustness with categorized ensemble of networks;improving real-time security screening;and coupling deep and handcrafted features to assess smile genuineness.
The review of "real-timevideo Object Detection using Deep Learning"provides an extensive analysis of the state-of-the-art in deep learning-powered real-timevideo object recognition systems. It examines the...
详细信息
The way to monitor the safety of infrastructure facilities such as power towers by human beings on the ground faces great risks under extreme environmental and climatic conditions. Therefore, automatic, real-time and ...
详细信息
ISBN:
(纸本)9789819772315;9789819772322
The way to monitor the safety of infrastructure facilities such as power towers by human beings on the ground faces great risks under extreme environmental and climatic conditions. Therefore, automatic, real-time and long-term monitoring of power towers in remote areas in the field through sensors, network communication and other technologies is the trend of today's technology development. However, when the real-time monitoring of high-definition images are captured by the camera and sent to the server for subsequent processing and analysis, the sheer volume of real-timeimage data causes pressure on the transmission network and the server side. Considering that in the real-time application of remote monitoring technology, the monitoring data obtained from sensors has redundant information, such as similar structure and repetitive background. We only need to extract the image data of the object of interest and compress it before transmission, therefore the image data is significantly transmitted to the server side, improving the efficiency of both network transmission and data processing. In this paper, we propose a learned image compression model by integrating a ResNet50 model and a Transformer-CNN-based network to reduce the image data that needs to be transmitted through the network and processed on the server side. The real-timeimage data is first sent to the ResNet50 model to extract objects of interest, which are then compressed by the Transformer-CNN network to realize remote monitoring by Learned image Compression (LIC) methods and communication techniques. Experimental results based on datasets collected in real-world scenarios indicate that the proposed solution effectively improves the compression performance compared to state-of-the-art methods. The average improvements in PSNR and MS-SSIM metrics are over 30%.
This paper presents a novel reconstruction algorithm for video Snapshot Compressive Imaging (SCI). Inspired by recent research works on Transformers and Self-Attention mechanism in computer vision, we propose the firs...
详细信息
ISBN:
(纸本)9781665469647
This paper presents a novel reconstruction algorithm for video Snapshot Compressive Imaging (SCI). Inspired by recent research works on Transformers and Self-Attention mechanism in computer vision, we propose the first video SCI reconstruction algorithm built upon Transformers to capture long-range spatio-temporal dependencies enabling the deep learning of feature maps. Our approach is based on a Spatiotemporal Convolutional Multi-head Attention (ST-ConvMHA) which enable to exploit the spatial and temporal information of the video scenes instead of using fully-connected attention layers. To evaluate the performances of our approach, we train our algorithm on DAVIS2017 dataset and we test the trained models on six benchmark datasets. The obtained results in terms of PSNR, SSIM and especially reconstruction time prove the ability of using our reconstruction approach for real-time applications. We truly believe that our research will motivate future works for more video reconstruction approaches.
Traditional imageprocessing mainly includes image digitization, dynamic range expansion, enhancement and denoising, compression coding, image feature extraction and so on. Moving target detection is an important subj...
详细信息
We propose two methods to improve the quality of extracted speech signals utilizing the out-of-focus areas in video captured with a rolling-shutter camera. A rolling-shutter camera exposes and reads pixels from the to...
详细信息
ISBN:
(纸本)9798350300673
We propose two methods to improve the quality of extracted speech signals utilizing the out-of-focus areas in video captured with a rolling-shutter camera. A rolling-shutter camera exposes and reads pixels from the top row of an image to the bottom, but when capturing an object that is vibrating due to speech, image distortion occurs due to the different exposure start times. The conventional method extracts a speech signal from a phase variation calculated from an image distortion. Here, we consider a case where out-of-focus areas arise in the captured video. In this case, the phase variation is not calculated correctly, which is expected to cause quality degradation of the speech signal extracted from the video. Our first proposed method weights the phase variation and the second one removes the out-of-focus areas. Experimental results show that our first method improves the quality of extracted speech signal and the second one reduces the time complexity.
Aerial image classification is essential to intelligent surveillance and monitoring systems. Traditional computer vision methods either uses computational offloading to high-end servers or edge devices. However, unman...
详细信息
ISBN:
(纸本)9798350367393;9798350367386
Aerial image classification is essential to intelligent surveillance and monitoring systems. Traditional computer vision methods either uses computational offloading to high-end servers or edge devices. However, unmanned aerial vehicles (UAVs) platforms have resource and power constraints. Aerial image classification is complicated and less-expensive UAVs lack processing power and cameras. Even with large-scale computing environments, methods for classifying images are difficult to apply to aerial imagery. We propose TinyAerialNet leveraging TinyML for real-time inference on a resource-constrained ESP32 CAM. The model tested on AIDER dataset, achieves 88% on-device accuracy in the micro-controller with 103.9 KB RAM and 850 milliseconds for inference.
Stereoscopic video conferencing is still challenging due to the need to compress stereo RGB-D video in real-time. Though hardware implementations of standard video codecs such as H.264 / AVC and HEVC are widely availa...
详细信息
ISBN:
(纸本)9798350365474
Stereoscopic video conferencing is still challenging due to the need to compress stereo RGB-D video in real-time. Though hardware implementations of standard video codecs such as H.264 / AVC and HEVC are widely available, they are not designed for stereoscopic videos and suffer from reduced quality and performance. Specific multiview or 3D extensions of these codecs are complex and lack efficient implementations. In this paper, we propose a new approach to upgrade a 2D video codec to support stereo RGB-D video compression, by wrapping it with a neural pre- and post-processor pair. The neural networks are end-to-end trained with an image codec proxy, and shown to work with a more sophisticated video codec. We also propose a geometry-aware loss function to improve rendering quality. We train the neural pre- and post-processors on a synthetic 4D people dataset, and evaluate it on both synthetic and real-captured stereo RGB-D videos. Experimental results show that the neural networks generalize well to unseen data and work out-of-box with various video codecs. Our approach saves about 30% bit-rate compared to a conventional video coding scheme and MV-HEVC at the same level of rendering quality from a novel view, without the need of a task-specific hardware upgrade.
暂无评论