Brain image segmentation is one of the most important parts of clinical diagnostic tools. Brain images mostly contain noise, inhomogeneity and sometimes deviation. Therefore, accurate segmentation of brain images is a very difficult task. However, the process of accurate segmentation of these images is very important and crucial for a correct diagnosis by clinical tools. We presented a review of the methods used in brain segmentation. The review covers imaging modalities, magnetic resonance imaging and methods for noise reduction, inhomogeneity correction and segmentation. We conclude with a discussion on the trend of future research in brain segmentation.

Breast cancer is the most common form of cancer among women worldwide. Early detection of breast cancer can increase treatment options and patients' survivability. Mammography is the gold standard for breast imaging and cancer detection. However, due to some limitations of this modality such as low sensitivity especially in dense breasts, other modalities like ultrasound and magnetic resonance imaging are often suggested to achieve additional information. Recently, computer-aided detection or diagnosis (CAD) systems have been developed to help radiologists in order to increase diagnosis accuracy. Generally, a CAD system consists of four stages: (a) preprocessing, (b) segmentation of regions of interest, (c) feature extraction and selection, and finally (d) classification. This paper presents the approaches which are applied to develop CAD systems on mammography and ultrasound images. The performance evaluation metrics of CAD systems are also reviewed.

Reliable feature correspondence between frames is a critical step in visual odometry (VO) and visual simultaneous localization and mapping (V-SLAM) algorithms. In comparison with existing VO and V-SLAM algorithms, semi-direct visual odometry (SVO) has two main advantages that lead to stateof-the-art frame rate camera motion estimation: direct pixel correspondence and efficient implementation of probabilistic mapping method. This paper improves the SVO mapping by initializing the mean and the variance of the depth at a feature location according to the depth prediction from a singleimage depth prediction network. By significantly reducing the depth uncertainty of the initialized map point (i.e., small variance centred about the depth prediction), the benefits are twofold: reliable feature correspondence between views and fast convergence to the true depth in order to create new map points. We evaluate our method with two outdoor datasets: KITTI dataset and Oxford Robotcar dataset. The experimental results indicate that the improved SVO mapping results in increased robustness and camera tracking accuracy.

No abstract

This paper proposes a fast method for car-license-plate detection (CLPD) and presents three main contributions. The first contribution is that we propose a fast vertical edge detection algorithm (VEDA) based on the contrast between the grayscale values, which enhances the speed of the CLPD method. After binarizing the input image using adaptive thresholding (AT), an unwanted-line elimination algorithm (ULEA) is proposed to enhance the image, and then, the VEDA is applied. The second contribution is that our proposed CLPD method processes very-low-resolution images taken by a web camera. After the vertical edges have been detected by the VEDA, the desired plate details based on color information are highlighted. Then, the candidate region based on statistical and logical operations will be extracted. Finally, an LP is detected. The third contribution is that we compare the VEDA to the Sobel operator in terms of accuracy, algorithm complexity, and processing time. The results show accurate edge detection performance and faster processing than Sobel by five to nine times. In terms of complexity, a big-O-notation module is used and the following result is obtained: The VEDA has less complexity by K2 times, whereas K2 represents the mask size of Sobel. Results show that the computation time of the CLPD method is 47.7 ms, which meets the real-time requirements.