Within the field of computer vision, change detection algorithms aim at automatically detecting significant changes occurring in a scene by analyzing the sequence of frames in a video stream. In this paper we investigate how state-of-the-art change detection algorithms can be combined and used to create a more robust algorithm leveraging their individual peculiarities. We exploited genetic programming (GP) to automatically select the best algorithms, combine them in different ways, and perform the most suitable post-processing operations on the outputs of the algorithms. In particular, algorithms’ combination and post-processing operations are achieved with unary, binary and ${n}$ -ary functions embedded into the GP framework. Using different experimental settings for combining existing algorithms we obtained different GP solutions that we termed In Unity There Is Strength . These solutions are then compared against state-of-the-art change detection algorithms on the video sequences and ground truth annotations of the ChangeDetection.net 2014 challenge. Results demonstrate that using GP, our solutions are able to outperform all the considered single state-of-the-art change detection algorithms, as well as other combination strategies. The performance of our algorithm are significantly different from those of the other state-of-the-art algorithms. This fact is supported by the statistical significance analysis conducted with the Friedman test and Wilcoxon rank sum post-hoc tests.

Combination of Video Change Detection Algorithms by Genetic Programming

A new scheme for nine-level voltage space-vector generation for medium-voltage induction motor (IM) drives with open-end stator winding is presented in this paper. The proposed nine-level power converter topology consists of two conventional three-phase two-level voltage source inverters powered by isolated dc sources and six floating-capacitor-connected $H$ -bridges. The $H$ -bridge capacitor voltages are effectively maintained at the required asymmetrical levels by employing a space vector modulation (SVPWM) based control strategy. An interesting feature of this topology is its ability to function in five- or three-level mode, in the entire modulation range, at full-power rating, in the event of any failure in the $H$ -bridges. This feature significantly improves the reliability of the proposed drive system. Each leg of the three-phase two-level inverters used in this topology switches only for a half cycle of the reference voltage waveform. Hence, the effective switching frequency is reduced by half, resulting in switching loss reduction in high-voltage devices. The transient as well as the steady-state performance of the proposed nine-level inverter-fed IM drive system is experimentally verified in the entire modulation range including the overmodulation region.

A Nine-Level Inverter Topology for Medium-Voltage Induction Motor Drive With Open-End Stator Winding

Glaucoma is a group of eye diseases which can cause vision loss by damaging the optic nerve. Early glaucoma detection is key to preventing vision loss yet there is a lack of noticeable early symptoms. Colour fundus photography allows the optic disc (OD) to be examined to diagnose glaucoma. Typically, this is done by measuring the vertical cup-to-disc ratio (CDR); however, glaucoma is characterised by thinning of the rim asymmetrically in the inferior-superior-temporal-nasal regions in increasing order. Automatic delineation of the OD features has potential to improve glaucoma management by allowing for this asymmetry to be considered in the measurements. Here, we propose a new deep-learning-based method to segment the OD and optic cup (OC). The core of the proposed method is DenseNet with a fully-convolutional network, whose symmetric U-shaped architecture allows pixel-wise classification. The predicted OD and OC boundaries are then used to estimate the CDR on two axes for glaucoma diagnosis. We assess the proposed method’s performance using a large retinal colour fundus dataset, outperforming state-of-the-art segmentation methods. Furthermore, we generalise our method to segment four fundus datasets from different devices without further training, outperforming the state-of-the-art on two and achieving comparable results on the remaining two.

Dense Fully Convolutional Segmentation of the Optic Disc and Cup in Colour Fundus for Glaucoma Diagnosis

This paper presents an application for counting people through a single fixed camera. This system performs the count distinction between input and output of people moving through the supervised area. The counter requires two steps: detection and tracking. The detection is based on finding people's heads through preprocessed image correlation with several circular patterns. Tracking is made through the application of a Kalman filter to determine the trajectory of the candidates. Finally, the system updates the counters based on the direction of the trajectories. Different tests using a set of real video sequences taken from different indoor areas give results ranging between 87% and 98% accuracies depending on the volume of flow of people crossing the counting zone. Problematic situations, such as occlusions, people grouped in different ways, scene luminance changes, etc., were used to validate the performance of the system.

/pdf/directional-people-counter-based-on-head-tracking-15owg8oovc.pdf

Directional People Counter Based on Head Tracking

Survey on segmentation and classification approaches of optic cup and optic disc for diagnosis of glaucoma

Fuzzy c-Means (FCM) for Optic Cup and Disc Segmentation with Morphological Operation

This paper studies the application of the Adaptive Neuro-Fuzzy Inference System (ANFIS) for segmentation of brain abnormality in MRI images. Segmentation of MRI image is an important part of brain imaging research. In this study, 150 MRI images were used as testing data for the system. The data was created by combining the shapes and size of various abnormalities and pasting it onto normal brain image. Several types of backgrounds were tested — low, medium and high grey levels. The experimental results show good segmentation for medium and low background levels value for both light and dark abnormality levels over different backgrounds.

Adaptive Neuro-Fuzzy Inference System for brain abnormality segmentation

This paper presents the enhancement capability of adaptive histogram equalization (AHE) on the soft tissue lateral neck radiograph for suspected fish bone ingestion. Embedded fish bone lodge in the throat is not easily visible in unprocessed plain radiograph. Serious complication may cause perforation of the lodged and inflammation that can progress to abscess. Forty X-ray images of 23 male and 17 female patients between the ages of 6 to 72 years from different ethnic groups were collected from Hospital Sungai Buloh and Pusat Perubatan Hospital Universiti Kebangsaan Malaysia (PPUKM). Due to the high resolution, the images were crop before being processed using adaptive histogram equalization. The quality of the image was assessed and evaluated during pre and post processing by the radiologists. The result shows AHE as a promising contrast enhancement for detection of fish bone in soft tissue at the lateral neck radiographs.

Fish Bone Impaction Using Adaptive Histogram Equalization (AHE)

This paper presents the performance evaluation between IPv4 and IPv6 with Linux MPLS tunnel. MPLS Linux tunneling is used to transport IPv6 data stream over IPv4 network for interoperable IPv4 and IPv6 deployment. The performance metrics such as jitter, datagram/packet loss and bandwidth were measured in both TCP and UPD traffic flow. The experiment test-bed was constructed using virtual machine tool for simulating the IPv4, IPv6, and the MPLS tunneling mechanism. This study revealed that IPv6 offered better performance than IPv4 in almost all of the testing except in TCP transfer. In addition, MPLS tunnel improved performance in packet transfer for TCP transfer, UDP transfer and bandwidth testing.

Performance evaluation between IPv4 and IPv6 on MPLS Linux platform

This paper reviews 80 out of 130 bio-inspired Algorithm (BIAs) researches published in google scholar and IEEExplore between the periods of 1995 to 2010 used to solve image processing problems. BIAs has been successfully applied in many sciences, medical and engineering fields. The evolving, dynamic and meta-heuristics nature of BIAs makes it more robust, accurate and efficient in solving image processing problems. However finding the appropriate BIAs that matches the problem at hand is a tedious and difficult task. The BIAs investigated in this study are Genetic Algorithms, Evolutionary strategies, Genetic programming, memetic algorithms, swarm intelligence and artificial immune system. The publications are categorized by year of publication, by specific BIAs and by application. The statistics shows exponential increases in the application of BIAs to solve image processing problems and some algorithms have yet to be explored.

Noorhayati Mohamed Noor

Papers

Fuzzy c-Means (FCM) for Optic Cup and Disc Segmentation with Morphological Operation

Adaptive Neuro-Fuzzy Inference System for brain abnormality segmentation

Fish Bone Impaction Using Adaptive Histogram Equalization (AHE)

Performance evaluation between IPv4 and IPv6 on MPLS Linux platform

A review of bio-inspired algorithms as image processing techniques