Enhancing YOLO deep networks for the detection of license plates in complex scenes

Al-qudah, Rabiah; Suen, Ching Y.

doi:10.1145/3368691.3368712

Cited by 5 publications

(3 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…By integrating the indirect and direct detection branches, we develop an end-to-end trainable network for license plate detection, which can effectively detect the small-sized license plate and accurately localize the multidirectional license plate in real applications. Combining Equations (1), (7), and (9), the loss of the whole network is shown in Equation 10, where α and β are simply set to 1 to balance these loss terms. Figure 4 illustrates the loss changes during training, including L 1 and L 2 of the indirect detection branch as well as L 3 of the direct detection branch.…”

Section: End-to-end Trainable Detection Networkmentioning

confidence: 99%

“…Yuan et al [ 6 ] apply dense filters to extract all the possible candidate LP regions and then preserve true positive LPs using a cascaded classifier. Rabiah et al [ 7 ] propose a YOLO-inspired adaptive solution with optimized parameters to enhance LPD performance. In literature [ 8 ], the license plate features from the bottom and high levels of the CNN network are extracted and integrated to achieve precise and real-time detection.…”

Section: Related Workmentioning

confidence: 99%

“…Recent LPD methods can be roughly divided into direct and indirect ways. Direct methods directly localize the license plate in the input image with handcrafted features [ 1 , 2 , 3 , 4 ], deep-learning features [ 5 , 6 , 7 , 8 , 9 ], or license plate recognition system [ 10 , 11 ]. However, detecting small-sized license plates is challenging since they only occupy a relatively small area in the whole image.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Scale-Invariant Multidirectional License Plate Detection with the Network Combining Indirect and Direct Branches

Chen

Liu

et al. 2021

Sensors

View full text Add to dashboard Cite

As the license plate is multiscale and multidirectional in the natural scene image, its detection is challenging in many applications. In this work, a novel network that combines indirect and direct branches is proposed for license plate detection in the wild. The indirect detection branch performs small-sized vehicle plate detection with high precision in a coarse-to-fine scheme using vehicle–plate relationships. The direct detection branch detects the license plate directly in the input image, reducing false negatives in the indirect detection branch due to the miss of vehicles’ detection. We propose a universal multidirectional license plate refinement method by localizing the four corners of the license plate. Finally, we construct an end-to-end trainable network for license plate detection by combining these two branches via post-processing operations. The network can effectively detect the small-sized license plate and localize the multidirectional license plate in real applications. To our knowledge, the proposed method is the first one that combines indirect and direct methods into an end-to-end network for license plate detection. Extensive experiments verify that our method outperforms the indirect methods and direct methods significantly.

show abstract

Section: End-to-end Trainable Detection Networkmentioning

confidence: 99%