No abstract

No abstract

Noise introduced during capture and transmission is inevitable for natural images generated by Complementary Metal-Oxide-Semiconductor (CMOS) sensors, including quantization error during digitalization, transmission disturbance and other sources of noise. To process natural images from a CMOS sensor, a hybrid filter combining Pulse Coupled Neural Network (PCNN), median filter and Wiener filter is proposed in this paper. First, salt-and-pepper noise is located via PCNN, and processed by a median filter. Then, Gaussian noise is removed by a self-adaptive Wiener filter. Simulation results indicated that compared to other methods (hybrid filter containing median and Wiener filter, hybrid filter containing median and wavelet filter), the hybrid filter with PCNN demonstrates better performance in the preservation of image detail and edge in the premise of similar Signal-Noise Ratios (SNRs).

In order to solve the problem of algorithm robustness caused by scale change and imbalanced distribution of classes in the scene text detection task, we propose a new two-stream fusion framework TSFnet. It is constructed by the Detection Stream, the Judge Stream and the Merge output algorithm. In the Detection Stream, we propose a loss balance factor (LBF), which is used to optimize region proposal network. Then, the Regression-net and the Segmentation-net are used to predict text global segmentation map and its corresponding coordinates probability score. In Judge Stream, we use the feature pyramid network to extract the Judge map. In the process, the LBF is calculated to support the Detection Stream. Finally, we design a novel algorithm to fuse the outputs of the two-stream, and the precise position of the text is localized. Extensive experiments are conducted on the ICDAR 2015 and the ICDAR2017-MLT standard datasets. The results demonstrate that the framework performance is comparable with the sate-of-the-art approaches.