Convolutional neural networks (CNNs) have achieved considerable success across a spectrum of computervision tasks, with applications ranging from healthcare to automated driving. Recent literature has also explored i...
详细信息
ISBN:
(纸本)9798400710810
Convolutional neural networks (CNNs) have achieved considerable success across a spectrum of computervision tasks, with applications ranging from healthcare to automated driving. Recent literature has also explored its potential utility in trading and risk management within the finance industry. Despite their versatility, CNNs are substantially constrained by their data-hungry nature. The lack of well-labeled image datasets poses a major challenge to the widespread adoption of CNNs in financial machine learning research across academia and industry. To address these concerns, this work presents Generative-CNN, a novel approach that utilizes a generative adversarial network (GAN) to synthetically generate images to enhance the performance of a CNN with applications in trading.
In the present research work, a system is developed that can detect objects in real-time using a combination of the ESP32 CAM module and Python programming. The goal was to show how affordable hardware and free softwa...
详细信息
ISBN:
(纸本)9798331540661;9798331540678
In the present research work, a system is developed that can detect objects in real-time using a combination of the ESP32 CAM module and Python programming. The goal was to show how affordable hardware and free software can be used to make a system that recognizes objects quickly and accurately. By using computervision and machine learning tricks, the proposed system can figure out the different objects with great precision. The setup process involves putting some code onto the ESP32 CAM module, finding its IP address, and then making it work smoothly with Python. The proposed system was tested in different situations, like watching for things in surveillance, making tasks easier with automation, and helping out in assistive technologies.
We propose an approach to domain adaptation for semantic segmentation that is both practical and highly accurate. In contrast to previous work, we abandon the use of computationally involved adversarial objectives, ne...
详细信息
ISBN:
(纸本)9781665445092
We propose an approach to domain adaptation for semantic segmentation that is both practical and highly accurate. In contrast to previous work, we abandon the use of computationally involved adversarial objectives, network ensembles and style transfer. Instead, we employ standard data augmentation techniques photometric noise, flipping and scaling and ensure consistency of the semantic predictions across these image transformations. We develop this principle in a lightweight self-supervised framework trained on co-evolving pseudo labels without the need for cumbersome extra training rounds. Simple in training from a practitioner's standpoint, our approach is remarkably effective. We achieve significant improvements of the state-of-the-art segmentation accuracy after adaptation, consistent both across different choices of the backbone architecture and adaptation scenarios.
The paper presents the research of computational complexity optimization of a virtual detector method for video-based vehicle recognition. This method enables a more flexible and adaptive analytical research and optim...
详细信息
In this paper we reformulate few-shot classification as a reconstruction problem in latent space. The ability of the network to reconstruct a query feature map from support features of a given class predicts membershi...
详细信息
ISBN:
(纸本)9781665445092
In this paper we reformulate few-shot classification as a reconstruction problem in latent space. The ability of the network to reconstruct a query feature map from support features of a given class predicts membership of the query in that class. We introduce a novel mechanism for few-shot classification by regressing directly from support features to query features in closed form, without introducing any new modules or large-scale learnable parameters. The resulting Feature Map Reconstruction Networks are both more performant and computationally efficient than previous approaches. We demonstrate consistent and substantial accuracy gains on four fine-grained benchmarks with varying neural architectures. Our model is also competitive on the non-fine-grained mini-ImageNet and tiered-ImageNet benchmarks with minimal bells and whistles.
Blind people are incapable of seeing, which is crucial for daily life. Blind people have limited autonomy due to their lack of eyesight. There are many methods for helping blind people navigate that are based on RFID,...
详细信息
A recurrent structure is a popular framework choice for the task of video super-resolution. The state-of-the-art method BasicVSR adopts bidirectional propagation with feature alignment to effectively exploit informati...
详细信息
This paper presents the development of a portable and interactive musical instrument using edge devices and various sensors. The goal is to create a versatile device that allows users to play different notes and melod...
详细信息
This paper presents an AI-based classroom monitoring system utilizing computervision and machine learning. The system employs Multi-Task Cascaded Convolutional Neural Networks (MTCNN) for face detection, FaceNet for ...
详细信息
ISBN:
(纸本)9798350308266;9798350308259
This paper presents an AI-based classroom monitoring system utilizing computervision and machine learning. The system employs Multi-Task Cascaded Convolutional Neural Networks (MTCNN) for face detection, FaceNet for facial recognition, and a CNN model trained on the Facial Expression recognition (FER) 2013 dataset for emotional analysis. A novel heterogeneous approach combines Field Programmable Gate Arrays (FPGA) and a central processor to overcome the limitations of BRAM and complex computation constraints. Evaluation in real-world classrooms yielded a promising 70% accuracy in emotion detection, marking a significant stride in the field. This research not only advances AI-based monitoring systems but also indicates potential applications in surveillance and security.
Image deblurring and super-resolution (SR) are computervision tasks aiming to restore image detail and spatial scale, respectively. Besides, only a few recent works of literature contribute to this task, as conventio...
详细信息
ISBN:
(纸本)9781665448994
Image deblurring and super-resolution (SR) are computervision tasks aiming to restore image detail and spatial scale, respectively. Besides, only a few recent works of literature contribute to this task, as conventional methods deal with SR or deblurring separately. We focus on designing a novel Pixel-Guided dual-branch attention network (PDAN) that handles both tasks jointly to address this issue. Then, we propose a novel loss function better focus on large and medium range errors. Extensive experiments demonstrated that the proposed PDAN with the novel loss function not only generates remarkably clear HR images and achieves compelling results for joint image deblurring and SR tasks. In addition, our method achieves second place in NTIRE 2021 Challenge on track 1 of the Image Deblurring Challenge.
暂无评论