As the effectiveness of an early detection of breast cancer using the mammography method alone is uncertain, it is crucial to provide an alternative method instead. This paper analyzes two optical flow algorithms utilizing a gradient method to aid current imaging techniques for a potential alternative method in aiding early breast cancer detection. The gradient method is a cost effective method that has the potential to be a mass screening method for this purpose. This paper compares two optical flow algorithms that are capable to detect the motion of breast tumor on B-mode ultrasound images. An analysis of 2D images of breast cancer lesions are compared using two gradient optical flow algorithms: Horn & Schunck and Lucas & Kanade. Both algorithms successfully show the direction of the tumor motion. However, while Lucas & Kanade can handle the short motion displacement of the tumor on the tested ultrasound images, Horn & Shunck failed to do so. This implies that the Lucas & Kanade algorithm is potentially more effective in handling ultrasound images of breast tumor. The results obtained showed that the Lucas&Kanade give better accuracy compared to Horn&Schunk.

Motion segmentation is a fundamental step for vehicle detection especially in urban traffic surveillance systems. Temporal frame differencing is the simplest and fastest technique that is used to identify foreground moving vehicles from static background scene. Conventional techniques utilize background modelling and subtraction, which involves poor adaptation under slow or temporarily stopped vehicles. To address this problems cumulative frame differencing (CFD) is proposed. Dynamic threshold value based on the standard deviation of CFD is used to estimate global variance of the motion accumulated variations of pixel intensity. The tests of the proposed technique achieve robust and accurate vehicle segmentation, which improves detection of slow motion, temporary and long term stopped vehicles, moreover, it enables the real-time capability.

Indoor pedestrian navigation suffers from the unavailability of useful GNSS signals for navigation. Often a low-cost non-GNSS inertial sensor is used to navigate indoors. However, using only a low-cost inertial sensor for the system degrades its performance due to the low observability of errors affecting such low-cost sensors. Of particular concern is the heading drift error, caused primarily by the unobservability of z-axis gyro bias errors, which results in a huge positioning error when navigating for more than a few seconds. In this paper, the observability of this error is increased by proposing a method of rotating the inertial sensor on its y-axis. The results from a field trial for the proposed innovative method are presented. The method was performed by rotating the sensor mechanically-mounted on a shoe-on a single axis. The method was shown to increase the observability of z-axis gyro bias errors of a low-cost sensor. This is very significant because no other integrated measurements from other sensors are required to increase error observability. This should potentially be very useful for autonomous low-cost inertial pedestrian navigation systems that require a long period of navigation time.

Outdoor images that are captured in bad weather conditions have low contrast and infidelity colours. Under the turbid medium conditions such as haze, mist, fog and drizzle, the light which reaches to the sensor is attenuated by atmospheric particles. These atmospheric phenomena degrade the contrast intensity of outdoor images based on haze density. In this research, we present new method to improve both the intensity and fine details of outdoor scene images. The RGB (Red, Green and Blue) input image is converted to the HSI (Hue Saturation Intensity) colour space and the density of the haze is estimated. Then, we use Contrast Limited Adaptive Histogram Equalization (CLAHE) technique to enhance the degraded intensity based on the estimation of the density of the haze. Our method is effective in a wide range of weather conditions and under different levels of visibility.

The haze phenomenon exerts a degrading effect that decreases contrast and causes colour shifts in outdoor images. The presence of haze in digital images is bothersome, unpleasant, and, occasionally, even dangerous. The atmospheric light scattering (ALS) model is widely used to restore hazy images. In this model, two unknown parameters should be estimated: airlight and scene transmission, The quality of dehazed images depends considerably on the accuracy of both estimates. Classic methods typically determine airlight based on the brightest pixels in an image. However, in the traffic scene context, this estimate is compromised when other light sources, such as vehicle headlights from the opposite direction, are present. Transmission estimation is usually more complicated. Hence, the complexity of the overall dehazing process is dependent on this estimate. To address this issue, this study proposes a framework for constant-time airlight and linear transmission estimation. This framework consists of two methods: airlight by image integrals (ALII), which is utilized to estimate the airlight value in real time with high accuracy, and bounded transmission (BT), which is proposed for the linear and simplified estimation of transmission maps. To evaluate the proposed framework, three image datasets are used: (1) seven images that are gathered from the works of existing methods (called the global dataset); (2) the synthetic foggy road image database (FRIDA), which is a synthetically generated dataset for simulating different bad weather conditions; and (3) a dataset of images that were extracted from videos in Malaysia (IV-M), which consists of images that were extracted from traffic video sequences, which were captured in various weather conditions from 2014 to 2016 in Malaysia. Experimental results show that the proposed framework is at least seven times faster than existing methods. In addition, the ANOVA test proves that the quality of the dehazed images is statistically similar to or better than the image quality that was achieved using existing methods. (C) 2018 Elsevier Inc. All rights reserved.

Rain, fog and haze are natural phenomena that fade scenes, limit the visibility range, and cause shifts in colors. These phenomena also play a decisive role in determining the degree of reliability of many kinds of outdoor applications, such as aerial and satellite imaging, surveillance, and driver assistance systems. Thus, removing their effects from images/videos is very crucial. Due to its mathematically ill posed nature, enhancement process of rain, fog, and haze plagued images/videos is highly challenging. In this paper, we propose a fast yet robust technique to enhance the visibility of video frames using the dark channel prior combined with fuzzy logic-based technique. The dark channel prior is a statistical regularity of outdoor haze-free images based on the observation that most local patches in the haze-free images contain pixels which are dark in at least one color channel, where the fuzzy logic-based technique is used to map an input space to an output space using a collection of fuzzy membership functions and rules to decide softly in case of uncertainties. The combination of the dark channel and the fuzzy logic-based technique will produce high quality haze-free images in real-time. Furthermore, it will be combined with rules derived from the stable atmospheric scattering model and will yield a fast yet high quality enhancement results.

Enhancing images that are plagued with weather related conditions; such as haze, fog and rain poses a challenging problem due to its ill-posed nature, which means the unknowns that need to be found are more than the equations that we have. To address such challenges, a fast yet robust method is proposed in this paper where unknowns in the light scattering model are estimated based on physically sound assumptions. Light scattering model describes the formation of those phenomena in an image as a combination of airlight and the original scene, where this combination is controlled by how much transmission value present at the scene's point. The transmission value determines how much of the original scene's intensity were attenuated and how much airlight was added. The attenuation term of the light scattering model causes the reduction in contrast and the airlight term causes the effect of color shift. In this paper, Intensity Deterioration Ratio (IDR) and Saturation Deterioration Ratio (SDR) are proposed, where the former can be used to estimate the reduction of contrast and so gives a clue about the attenuation term, and the latter to estimate the reduction of chromaticity in a scene which gives a clue about the airlight term. IDR and SDR are therefore used to give us a new insight in using the light scattering model when enhancing images.

Shoe mounted Inertial Measurement Units (IMU) are often used for indoor pedestrian navigation systems. The presence of a zero velocity condition during the stance phase enables Zero Velocity Updates (ZUPT) to be applied regularly every time the user takes a step. Most of the velocity and attitude errors can be estimated using ZUPTs. However, good heading estimation for such a system remains a challenge. This is due to the poor observability of heading error for a low cost Micro-Electro-Mechanical (MEMS) IMU, even with the use of ZUPTs in a Kalman filter. In this paper, the same approach is adopted where a MEMS IMU is mounted on a shoe, but with additional constraints applied. The three constraints proposed herein are used to generate measurement updates for a Kalman filter, known as 'Heading Update', 'Zero Integrated Heading Rate Update' and 'Height Update'. The first constraint involves restricting heading drift in a typical building where the user is walking. Due to the fact that typical buildings are rectangular in shape, an assumption is made that most walking in this environment is constrained to only follow one of the four main headings of the building. A second constraint is further used to restrict heading drift during a non-walking situation. This is carried out because the first constraint cannot be applied when the user is stationary. Finally, the third constraint is applied to limit the error growth in height. An assumption is made that the height changes in indoor buildings are only caused when the user walks up and down a staircase. Several trials were shown to demonstrate the effectiveness of integrating these constraints for indoor pedestrian navigation. The results show that an average return position error of 4.62 meters is obtained for an average distance of 1557 meters using only a low cost MEMS IMU.