Simplified Neural Networks with Smart Detection for Road Traffic Sign Recognition

Yang, Wei; Luo, Chia Chun; Chung, Pau Choo; Yang, Jar Ferr

doi:10.1007/978-3-030-12388-8_17

Cited by 10 publications

(4 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The ability of the proposed method is demonstrated and compared with sophisticated and popular shallow and deep learning approaches [42][25] [24][26] [27] on a challenging noisy audio-visual speech processing task that uses video information from lip movements to selectively amplify speech signals heard in noisy environments. It is observed that MCC is able to remove background noise with better reconstruction than the state-of-the-art baseline shallow and deep learning algorithms.…”

Section: Methodsmentioning

confidence: 99%

“…It is observed that MCC is able to remove background noise with better reconstruction than the state-of-the-art baseline shallow and deep learning algorithms. For fair comparisons, both shallow and deep benchmark models have C3/attention blocks integrated [24][25][26] [27]. The C3 or cross-channel fusion is implemented through concatenation, addition, or multiplication using LIF-inspired point neural model.…”

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Unlocking the Potential of Two-Point Cells for Energy-Efficient and Resilient Training of Deep Nets

Adeel

Adetomi

Ahmed

et al. 2023

IEEE Trans. Emerg. Top. Comput. Intell.

View full text Add to dashboard Cite

Context-sensitive two-point layer 5 pyramidal cells (L5PCs) were discovered as long ago as 1999. However, the potential of this discovery to provide useful neural computation has yet to be demonstrated. Here we show for the first time how a transformative L5PCs-driven deep neural network (DNN), termed the multisensory cooperative computing (MCC) architecture, can effectively process large amounts of heterogeneous realworld audio-visual (AV) data, using far less energy compared to best available 'point' neuron-driven DNNs. A novel highlydistributed parallel implementation on a Xilinx UltraScale+ MPSoC device estimates energy savings up to 245759 × 50000 µJ (i.e., 62% less than the baseline model in a semi-supervised learning setup) where a single synapse consumes 8e −5 µJ. In a supervised learning setup, the energy-saving can potentially reach up to 1250x less (per feedforward transmission) than the baseline model. The significantly reduced neural activity in MCC leads to inherently fast learning and resilience against sudden neural damage. This remarkable performance in pilot experiments demonstrates the embodied neuromorphic intelligence of our proposed cooperative L5PC that receives input from diverse neighbouring neurons as context to amplify the transmission of most salient and relevant information for onward transmission, from overwhelmingly large multimodal information utilised at the early stages of on-chip training. Our proposed approach opens new cross-disciplinary avenues for future on-chip DNN training implementations and posits a radical shift in current neuromorphic computing paradigms.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Unlocking the Potential of Two-Point Cells for Energy-Efficient and Resilient Training of Deep Nets

Adeel

Adetomi

Ahmed

et al. 2023

IEEE Trans. Emerg. Top. Comput. Intell.

View full text Add to dashboard Cite

show abstract

“…Neighbors. [6] proposed an approach with shaped based detection algorithms to recognize traffic signs. The authors chose convolutional neural network for the purpose of classification.…”

Section: Related Workmentioning

confidence: 99%

Lightweight Residual Layers Based Convolutional Neural Networks for Traffic Sign Recognition

Mamatkulovich¹

2022

eijmrms

View full text Add to dashboard Cite

System for Traffic Sign Recognition and Classification is significantly important for especially traffic safety, traffic surveillance, artificial driver services and by all means, for self-driving cars. Traffic sign recognition plays an important role to tackle the traffic related obstacles. And, as traffic sign recognition is particularly applied to portable devices, lightweight models are essential aspect of the agenda. To overcome the mentioned problems, we propose lightweight convolutional neural networks with residual blocks based deep learning model for traffic recognition systems. We not only present the model efficiency but also show the several conducted experiments will well known deep CNN architectures over publicly available German traffic sign recognition benchmark. Our model showed 99.9 % accuracy by F-score, exceeding other models. At last, our model shows generally validity for traffic sign classification problem.

show abstract

“…La tecnología permite mejorar el tiempo de ejecución de las actividades que realiza el hombre, por ello la visión po r computador es un tema actualmente muy estudiado que ha permitido realizar tareas como la detección de rostros, conteo de vehículos, detección de letras y otras actividades, [9] . Para esta investigación se ha revisado bibliografía con referencia a identificación de señales de tránsito y en su mayoría tiene por objetivo mejorar las aplicaciones en vehículos inteligentes [7] , [11] , [25] . La metodología se ha dividido por lo general en dos etapas: la detección y el reconocimiento.…”

Section: Introductionunclassified

Detection and recognition of Peruvian warning traffic signals applied to road inventories-case Eduardo of Habich avenue

Ramos¹,

Flores²,

Valero³

et al. 2022

Proceedings of the 20th LACCEI International Multi-Conference for Engineering, Education and Technology: “Education, Research A

View full text Add to dashboard Cite

The purpose of this research is to obtain a database of Peruvian warning traffic signs and propose a tool to automate the road inventory process using image processing algorithms. The database consists of 2026 images of Peruvian warning traffic signs, to detect and recognize them on Av. Eduardo Habich located in Metropolitan Lima, also proposed the following methodology tha t is divided into two parts: The first part consists of collecting da ta in the field for the creation of the database; and the second part consists of the processing of information in the cabinet, where the detection and recognition algorithm for information processing is proposed. The detection stage consists of the use of color and sha p e filters, as well as the performance of two-color models, HSV a nd normalized RGB, for the characteristic yellow color of warning signs. The recognition stage consists of the use of supervised classification tools with the algorithm called support vector machines. Finally, with the development of this research, it was possible to obtain an algorithm that allows the detection of tra ffic signs with a recognition percentage of 62.5% and a solid database that can be fed back and give rise to future research in the automation of traffic signals. road inventories.

show abstract

Simplified Neural Networks with Smart Detection for Road Traffic Sign Recognition

Cited by 10 publications

References 23 publications

Unlocking the Potential of Two-Point Cells for Energy-Efficient and Resilient Training of Deep Nets

Unlocking the Potential of Two-Point Cells for Energy-Efficient and Resilient Training of Deep Nets

Lightweight Residual Layers Based Convolutional Neural Networks for Traffic Sign Recognition

Detection and recognition of Peruvian warning traffic signals applied to road inventories-case Eduardo of Habich avenue

Contact Info

Product

Resources

About