BLPnet: A new DNN model and Bengali OCR engine for Automatic Licence Plate Recognition

Onim, Md. Saif Hassan; Nyeem, Hussain; Roy, Koushik; Hasan, Mahmudul; Ishmam, Abtahi; Akif, Md. Akiful Hoque; Ovi, Tareque Bashar

doi:10.1016/j.array.2022.100244

Cited by 15 publications

(12 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…With bidirectional LSTM network, Zou et al [21] achieved an accuracy of 35%, precision of 10%, and IoU of 12% on dark images, while on daylight images, it achieved an accuracy of 87%, precision of 60%, and IoU of 71%. Onim et al [22]'s CNN-based BLPNet achieved an accuracy of 65%, precision of 27%, and IoU of 76% on dark images, while on daylight images, the model achieved an accuracy of 89%, precision of 87%, and IoU of 82%. The above-mentioned models exhibit remarkable performance when presented with daylight images.…”

Section: Comparative Analysismentioning

confidence: 99%

“…Zou et al [21] introduced a Bi-LSTM-based feature extraction model for licence plate recognition, achieving an accuracy of 97.8%, 96.5%, 80.7%, and 99.5% on the CCPD, PKUdata, CLPD, and AOLP datasets, respectively. Onim et al [22] proposed a CNN-based licence plate detection model that reported a mean accuracy of 95% on a customised dataset.…”

Section: Road-object Detection Modelsmentioning

confidence: 99%

“…Onim et al . [22] proposed a CNN‐based licence plate detection model that reported a mean accuracy of 95% on a customised dataset.…”

Section: Related Workmentioning

confidence: 99%

“…Recent vision-based techniques employing cameras with visible-light or infrared sensors have made significant progress in road-object detection and tracking. These techniques leverage object detection and tracking algorithms with Machine Learning (ML) to achieve near-human-level accuracy in detecting and localising objects on the road [13][14][15][16][17][18][19][20][21][22]. However, they are mostly limited to daylight conditions, ignoring lowlight or dark conditions.…”

mentioning

confidence: 99%

“…To test the localisation performance of vehicle images in extremely lowlight conditions, we developed a new dark version dataset from the popular vehicle datasets (i.e., PKUdata-cars [30], Stanford-cars [31], and Caltech-cars [32]). Finally, we analyse and demonstrate the potential of the proposed model with its impressive detection and localisation ability over the stateof-the-art vehicle detection models [17][18][19][20][21][22] in both the daylight and low-light conditions of a traffic scenario (see Section 4).…”

mentioning

confidence: 99%

See 4 more Smart Citations

Unleashing the power of generative adversarial networks: A novel machine learning approach for vehicle detection and localisation in the dark

Hassan Onim,

Nyeem,

Khan Arnob

et al. 2023

Cognitive Comp and Systems

Self Cite

View full text Add to dashboard Cite

Machine vision in low‐light conditions is a critical requirement for object detection in road transportation, particularly for assisted and autonomous driving scenarios. Existing vision‐based techniques are limited to daylight traffic scenarios due to their reliance on adequate lighting and high frame rates. This paper presents a novel approach to tackle this problem by investigating Vehicle Detection and Localisation (VDL) in extremely low‐light conditions by using a new machine learning model. Specifically, the proposed model employs two customised generative adversarial networks, based on Pix2PixGAN and CycleGAN, to enhance dark images for input into a YOLOv4‐based VDL algorithm. The model's performance is thoroughly analysed and compared against the prominent models. Our findings validate that the proposed model detects and localises vehicles accurately in extremely dark images, with an additional run‐time of approximately 11 ms and an accuracy improvement of 10%–50% compared to the other models. Moreover, our model demonstrates a 4%–8% increase in Intersection over Union (IoU) at a mean frame rate of 9 fps, which underscores its potential for broader applications in ubiquitous road‐object detection. The results demonstrate the significance of the proposed model as an early step to overcoming the challenges of low‐light vision in road‐object detection and autonomous driving, paving the way for safer and more efficient transportation systems.

show abstract

Section: Comparative Analysismentioning

confidence: 99%

Section: Road-object Detection Modelsmentioning

confidence: 99%

“…Onim et al . [22] proposed a CNN‐based licence plate detection model that reported a mean accuracy of 95% on a customised dataset.…”

Section: Related Workmentioning

confidence: 99%