Search

Printed bank cards number, which are usually black and not easily confused with the background, is easy to split and recognize; embossed bank card, the color of bumped card number is similar to background, characters difficult to split and identify. This paper proposes an end-to-end bank card number recognition algorithm, first uses CTPN network for text detection, then sends the test results into the DenseNet network to identify, finally the card number is output according to the text length. Experimental results show that the algorithm can achieve the automatic identification of the two types of bank card, with high recognition rate and accuracy.

We develop an image reconstruction algorithm in this paper using inverse transformation of principal component residual to automatically detect the tire defects in its X-ray image. A tire X-ray image to be inspected is first transformed to a new coordinate system, i.e. principal component space, in which the k major components and their corresponding eigenvectors represent the dominant normal textures of tires; Unlike traditional principal component analysis, we set the k major component eigenvalues as zeros, then the above principal component space matrix is called principal component residual in this paper, which stands for none-dominant textures of tires that is considered as abnormal or defect textures. We perform inverse transformation of this principal component residual and reconstruct the defect left only image. To localize the defect part, binarization operation is processed on the obtained left over image with upper and lower thresholds statistically. The proposed scheme can both reveal the defect locations and also present defect patch shapes for identifications. The experiment results show the high effectiveness of our scheme in detecting defects for tire X-ray images.

Pedestrian tracking is a critical problem in the field of computer vision. Particle filters have been proven to be very useful in pedestrian tracking for nonlinear and non-Gaussian estimation problems. However, pedestrian tracking in complex environment is still facing many problems due to changes of pedestrian postures and scale, moving background, mutual occlusion, and presence of pedestrian. To surmount these difficulties, this paper presents tracking algorithm of multiple pedestrians based on particle filters in video sequences. The algorithm acquires confidence value of the object and the background through extracting a priori knowledge thus to achieve multipedestrian detection; it adopts color and texture features into particle filter to get better observation results and then automatically adjusts weight value of each feature according to current tracking environment. During the process of tracking, the algorithm processes severe occlusion condition to prevent drift and loss phenomena caused by object occlusion and associates detection results with particle state to propose discriminated method for object disappearance and emergence thus to achieve robust tracking of multiple pedestrians. Experimental verification and analysis in video sequences demonstrate that proposed algorithm improves the tracking performance and has better tracking results.

The objective of this study is to improve the accuracy in tire defect classification with limited training samples under varying illuminations. We investigate an algorithm based on deep learning to achieve high accuracy with limited samples. First, image contrast normalizations and data augmentation were used to avoid overfitting problems of the network with a large number of parameters. Furthermore, multi-column CNN is proposed by combining several CNNs trained on differently preprocessed data into a multi-column CNN (MC-CNN), and then their predictions are averaged as the output of the proposed network. An average accuracy of 98.47% is achieved with the proposed CNN-based method. Experimental results show that our scheme receives satisfactory classification accuracy and outperforms state-of-the-art methods on the same tire defect dataset.

For inpainting of the flow-like textures with line-like structures images, we present an model based on tensor diffusion. The proposed model is to reconstruct a damaged original image by acting along the texture's direction and the vertical texture's direction which is determined by the texture's trend of the local image structure (This local image structure is measured by the so-called structure tensor). In order to guarantee better results, the intensity of the diffusion along the texture's direction should be much greater than that of the diffusion along the vertical texture's direction. The nonnegativity discretization method and optimized rotation invariance method is used to calculate the proposed model. It was shown in experiments that this model can get good results for inpainting the flow-like textures with line-like structures images that contain nicks or small broken areas.