Automatic Individual Pig Detection and Tracking in Pig Farms.

Individual pig detection and tracking is an important requirement in many video-based pig monitoring applications. However, it still remains a challenging task in complex scenes, due to problems of light fluctuation, similar appearances of pigs, shape deformations, and occlusions. In order to tackle these problems, we propose a robust on-line multiple pig detection and tracking method which does not require manual marking or physical identification of the pigs and works under both daylight and infrared (nighttime) light conditions. Our method couples a CNN-based detector and a correlation filter-based tracker via a novel hierarchical data association algorithm. In our method, the detector gains the best accuracy/speed trade-off by using the features derived from multiple layers at different scales in a one-stage prediction network. We define a tag-box for each pig as the tracking target, from which features with a more local scope are extracted for learning, and the multiple object tracking is conducted in a key-points tracking manner using learned correlation filters. Under challenging conditions, the tracking failures are modelled based on the relations between responses of the detector and tracker, and the data association algorithm allows the detection hypotheses to be refined; meanwhile the drifted tracks can be corrected by probing the tracking failures followed by the re-initialization of tracking. As a result, the optimal tracklets can sequentially grow with on-line refined detections, and tracking fragments are correctly integrated into respective tracks while keeping the original identifications. Experiments with a dataset captured from a commercial farm show that our method can robustly detect and track multiple pigs under challenging conditions. The promising performance of the proposed method also demonstrates the feasibility of long-term individual pig tracking in a complex environment and thus promises commercial potential.

PRaVDA: The first solid-state system for proton computed tomography.

PURPOSE: Proton CT is widely recognised as a beneficial alternative to conventional X-ray CT for treatment planning in proton beam radiotherapy. A novel proton CT imaging system, based entirely on solid-state detector technology, is presented. Compared to conventional scintillator-based calorimeters, positional sensitive detectors allow for multiple protons to be tracked per read out cycle, leading to a potential reduction in proton CT scan time. Design and characterisation of its components are discussed. An early proton CT image obtained with a fully solid-state imaging system is shown and accuracy (as defined in Section IV) in Relative Stopping Power to water (RSP) quantified. METHOD: A solid-state imaging system for proton CT, based on silicon strip detectors, has been developed by the PRaVDA collaboration. The system comprises a tracking system that infers individual proton trajectories through an imaging phantom, and a Range Telescope (RT) which records the corresponding residual energy (range) for each proton. A back-projection-then-filtering algorithm is used for CT reconstruction of an experimentally acquired proton CT scan. RESULTS: An initial experimental result for proton CT imaging with a fully solid-state system is shown for an imaging phantom, namely a 75 mm diameter PMMA sphere containing tissue substitute inserts, imaged with a passively-scattered 125 MeV beam. Accuracy in RSP is measured to be ⩽1.6% for all the inserts shown. CONCLUSIONS: A fully solid-state imaging system for proton CT has been shown capable of imaging a phantom with protons and successfully improving RSP accuracy. These promising results, together with system the capability to cope with high proton fluences (2×108 protons/s), suggests that this research platform could improve current standards in treatment planning for proton beam radiotherapy.

Supervised learning based multimodal MRI brain tumour segmentation using texture features from supervoxels.

BACKGROUND: Accurate segmentation of brain tumour in magnetic resonance images (MRI) is a difficult task due to various tumour types. Using information and features from multimodal MRI including structural MRI and isotropic (p) and anisotropic (q) components derived from the diffusion tensor imaging (DTI) may result in a more accurate analysis of brain images. METHODS: We propose a novel 3D supervoxel based learning method for segmentation of tumour in multimodal MRI brain images (conventional MRI and DTI). Supervoxels are generated using the information across the multimodal MRI dataset. For each supervoxel, a variety of features including histograms of texton descriptor, calculated using a set of Gabor filters with different sizes and orientations, and first order intensity statistical features are extracted. Those features are fed into a random forests (RF) classifier to classify each supervoxel into tumour core, oedema or healthy brain tissue. RESULTS: The method is evaluated on two datasets: 1) Our clinical dataset: 11 multimodal images of patients and 2) BRATS 2013 clinical dataset: 30 multimodal images. For our clinical dataset, the average detection sensitivity of tumour (including tumour core and oedema) using multimodal MRI is 86% with balanced error rate (BER) 7%; while the Dice score for automatic tumour segmentation against ground truth is 0.84. The corresponding results of the BRATS 2013 dataset are 96%, 2% and 0.89, respectively. CONCLUSION: The method demonstrates promising results in the segmentation of brain tumour. Adding features from multimodal MRI images can largely increase the segmentation accuracy. The method provides a close match to expert delineation across all tumour grades, leading to a faster and more reproducible method of brain tumour detection and delineation to aid patient management.

Automatic image quality assessment and measurement of fetal head in two-dimensional ultrasound image.

Owing to the inconsistent image quality existing in routine obstetric ultrasound (US) scans that leads to a large intraobserver and interobserver variability, the aim of this study is to develop a quality-assured, fully automated US fetal head measurement system. A texton-based fetal head segmentation is used as a prerequisite step to obtain the head region. Textons are calculated using a filter bank designed specific for US fetal head structure. Both shape- and anatomic-based features calculated from the segmented head region are then fed into a random forest classifier to determine the quality of the image (e.g., whether the image is acquired from a correct imaging plane), from which fetal head measurements [biparietal diameter (BPD), occipital-frontal diameter (OFD), and head circumference (HC)] are derived. The experimental results show a good performance of our method for US quality assessment and fetal head measurements. The overall precision for automatic image quality assessment is 95.24% with 87.5% sensitivity and 100% specificity, while segmentation performance shows 99.27% ([Formula: see text]) of accuracy, 97.07% ([Formula: see text]) of sensitivity, 2.23 mm ([Formula: see text]) of the maximum symmetric contour distance, and 0.84 mm ([Formula: see text]) of the average symmetric contour distance. The statistical analysis results using paired [Formula: see text]-test and Bland-Altman plots analysis indicate that the 95% limits of agreement for inter observer variability between the automated measurements and the senior expert measurements are 2.7 mm of BPD, 5.8 mm of OFD, and 10.4 mm of HC, whereas the mean differences are [Formula: see text], [Formula: see text], and [Formula: see text], respectively. These narrow 95% limits of agreements indicate a good level of consistency between the automated and the senior expert's measurements.

Automated brain tumour detection and segmentation using superpixel-based extremely randomized trees in FLAIR MRI.

PURPOSE: We propose a fully automated method for detection and segmentation of the abnormal tissue associated with brain tumour (tumour core and oedema) from Fluid- Attenuated Inversion Recovery (FLAIR) Magnetic Resonance Imaging (MRI). METHODS: The method is based on superpixel technique and classification of each superpixel. A number of novel image features including intensity-based, Gabor textons, fractal analysis and curvatures are calculated from each superpixel within the entire brain area in FLAIR MRI to ensure a robust classification. Extremely randomized trees (ERT) classifier is compared with support vector machine (SVM) to classify each superpixel into tumour and non-tumour. RESULTS: The proposed method is evaluated on two datasets: (1) Our own clinical dataset: 19 MRI FLAIR images of patients with gliomas of grade II to IV, and (2) BRATS 2012 dataset: 30 FLAIR images with 10 low-grade and 20 high-grade gliomas. The experimental results demonstrate the high detection and segmentation performance of the proposed method using ERT classifier. For our own cohort, the average detection sensitivity, balanced error rate and the Dice overlap measure for the segmented tumour against the ground truth are 89.48 %, 6 % and 0.91, respectively, while, for the BRATS dataset, the corresponding evaluation results are 88.09 %, 6 % and 0.88, respectively. CONCLUSIONS: This provides a close match to expert delineation across all grades of glioma, leading to a faster and more reproducible method of brain tumour detection and delineation to aid patient management.

A supervised texton based approach for automatic segmentation and measurement of the fetal head and femur in 2D ultrasound images.

This paper presents a supervised texton based approach for the accurate segmentation and measurement of ultrasound fetal head (BPD, OFD, HC) and femur (FL). The method consists of several steps. First, a non-linear diffusion technique is utilized to reduce the speckle noise. Then, based on the assumption that cross sectional intensity profiles of skull and femur can be approximated by Gaussian-like curves, a multi-scale and multi-orientation filter bank is designed to extract texton features specific to ultrasound fetal anatomic structure. The extracted texton cues, together with multi-scale local brightness, are then built into a unified framework for boundary detection of ultrasound fetal head and femur. Finally, for fetal head, a direct least square ellipse fitting method is used to construct a closed head contour, whilst, for fetal femur a closed contour is produced by connecting the detected femur boundaries. The presented method is demonstrated to be promising for clinical applications. Overall the evaluation results of fetal head segmentation and measurement from our method are comparable with the inter-observer difference of experts, with the best average precision of 96.85%, the maximum symmetric contour distance (MSD) of 1.46 mm, average symmetric contour distance (ASD) of 0.53 mm; while for fetal femur, the overall performance of our method is better than the inter-observer difference of experts, with the average precision of 84.37%, MSD of 2.72 mm and ASD of 0.31 mm.

Towards real-time VMAT verification using a prototype, high-speed CMOS active pixel sensor.

This work investigates the feasibility of using a prototype complementary metal oxide semiconductor active pixel sensor (CMOS APS) for real-time verification of volumetric modulated arc therapy (VMAT) treatment. The prototype CMOS APS used region of interest read out on the chip to allow fast imaging of up to 403.6 frames per second (f/s). The sensor was made larger (5.4 cm × 5.4 cm) using recent advances in photolithographic technique but retains fast imaging speed with the sensor's regional read out. There is a paradigm shift in radiotherapy treatment verification with the advent of advanced treatment techniques such as VMAT. This work has demonstrated that the APS can track multi leaf collimator (MLC) leaves moving at 18 mm s(-1) with an automatic edge tracking algorithm at accuracy better than 1.0 mm even at the fastest imaging speed. Evaluation of the measured fluence distribution for an example VMAT delivery sampled at 50.4 f/s was shown to agree well with the planned fluence distribution, with an average gamma pass rate of 96% at 3%/3 mm. The MLC leaves motion and linac pulse rate variation delivered throughout the VMAT treatment can also be measured. The results demonstrate the potential of CMOS APS technology as a real-time radiotherapy dosimeter for delivery of complex treatments such as VMAT.

Imaging of moving fiducial markers during radiotherapy using a fast, efficient active pixel sensor based EPID.

PURPOSE: The purpose of this work was to investigate the use of an experimental complementary metal-oxide-semiconductor (CMOS) active pixel sensor (APS) for tracking of moving fiducial markers during radiotherapy. METHODS: The APS has an active area of 5.4 × 5.4 cm and maximum full frame read-out rate of 20 frame s(-1), with the option to read out a region-of-interest (ROI) at an increased rate. It was coupled to a 4 mm thick ZnWO4 scintillator which provided a quantum efficiency (QE) of 8% for a 6 MV x-ray treatment beam. The APS was compared with a standard iViewGT flat panel amorphous Silicon (a-Si) electronic portal imaging device (EPID), with a QE of 0.34% and a frame-rate of 2.5 frame s(-1). To investigate the ability of the two systems to image markers, four gold cylinders of length 8 mm and diameter 0.8, 1.2, 1.6, and 2 mm were placed on a motion-platform. Images of the stationary markers were acquired using the APS at a frame-rate of 20 frame s(-1), and a dose-rate of 143 MU min(-1) to avoid saturation. EPID images were acquired at the maximum frame-rate of 2.5 frame s(-1), and a reduced dose-rate of 19 MU min(-1) to provide a similar dose per frame to the APS. Signal-to-noise ratio (SNR) of the background signal and contrast-to-noise ratio (CNR) of the marker signal relative to the background were evaluated for both imagers at doses of 0.125 to 2 MU. RESULTS: Image quality and marker visibility was found to be greater in the APS with SNR ∼5 times greater than in the EPID and CNR up to an order of magnitude greater for all four markers. To investigate the ability to image and track moving markers the motion-platform was moved to simulate a breathing cycle with period 6 s, amplitude 20 mm and maximum speed 13.2 mm s(-1). At the minimum integration time of 50 ms a tracking algorithm applied to the APS data found all four markers with a success rate of ≥92% and positional error ≤90 µm. At an integration time of 400 ms the smallest marker became difficult to detect when moving. The detection of moving markers using the a-Si EPID was difficult even at the maximum dose-rate of 592 MU min(-1) due to the lower QE and longer integration time of 400 ms. CONCLUSIONS: This work demonstrates that a fast read-out, high QE APS may be useful in the tracking of moving fiducial markers during radiotherapy. Further study is required to investigate the tracking of markers moving in 3D in a treatment beam attenuated by moving patient anatomy. This will require a larger sensor with ROI read-out to maintain speed and a manageable data-rate.

Multitraining support vector machine for image retrieval.

Relevance feedback (RF) schemes based on support vector machines (SVMs) have been widely used in content-based image retrieval (CBIR). However, the performance of SVM-based RF approaches is often poor when the number of labeled feedback samples is small. This is mainly due to 1) the SVM classifier being unstable for small-size training sets because its optimal hyper plane is too sensitive to the training examples; and 2) the kernel method being ineffective because the feature dimension is much greater than the size of the training samples. In this paper, we develop a new machine learning technique, multitraining SVM (MTSVM), which combines the merits of the cotraining technique and a random sampling method in the feature space. Based on the proposed MTSVM algorithm, the above two problems can be mitigated. Experiments are carried out on a large image set of some 20,000 images, and the preliminary results demonstrate that the developed method consistently improves the performance over conventional SVM-based RFs in terms of precision and standard deviation, which are used to evaluate the effectiveness and robustness of a RF algorithm, respectively.

Estimation of protein secondary structure from synchrotron radiation circular dichroism spectra.

A new approach to estimate the fraction of secondary structures fractions from synchrotron radiation circular dichroism (SRCD) spectra is presented. The protein SRCD spectra are first approximated using radial basis function networks (RBFN) and the resulting set is used to train a self-organising map (SOM). Thus the data are arranged in a two-dimensional map in such a way that most similar proteins are close to each other and vice versa. Estimation of the parallel and antiparallel beta sheets is discussed. The number of spectra in the training set is twenty four proteins and the protein under examination is also included in the set. Estimation results shows improvements compared with previous methods such as K2D and SOMCD.

Image denoising using self-organizing map-based nonlinear independent component analysis.

This paper proposes the use of self-organizing maps (SOMs) to the blind source separation (BSS) problem for nonlinearly mixed signals corrupted with multiplicative noise. After an overview of some signal denoising approaches, we introduce the generic independent component analysis (ICA) framework, followed by a survey of existing neural solutions on ICA and nonlinear ICA (NLICA). We then detail a BSS method based on SOMs and intended for image denoising applications. Considering that the pixel intensities of raw images represent a useful signal corrupted with noise, we show that an NLICA-based approach can provide a satisfactory solution to the nonlinear BSS (NLBSS) problem. Furthermore, a comparison between the standard SOM and a modified version, more suitable for dealing with multiplicative noise, is made. Separation results obtained from test and real images demonstrate the feasibility of our approach.

Basis function models of the CMAC network.

An interpretation of the Cerebellar Model Articulation Controller (CMAC) network as a member of the General Memory Neural Network (GMNN) architecture is presented. The usefulness of this approach stems from the fact that, within the GMNN formalism, CMAC can be treated as a particular form of a basis function network, where the basis function is inherently dependent on the type of input quantization present in the network mapping. Furthermore, considering the relative regularity of input-space quantization performed by CMAC, we are able to derive an expected (or average) form of the basis function characteristic of this network. Using this basis form, it is possible to create basis-functions models of CMAC mapping, as well as to gain more insight into its performance. The developments are supported by numerical simulations.

N-tuple Regression Network.

N-tuple neural networks (NTNNs) have been successfully applied to both pattern recognition and function approximation tasks. Their main advantages include a single layer structure, capability of realizing highly non-linear mappings and simplicity of operation. In this work a modification of the basic network architecture is presented, which allows it to operate as a non-parametric kernel regression estimator. This type of network is inherently capable of approximating complex probability density functions (pdfs) and, in the limiting sense, deterministic arbitrary function mappings. At the same time, the regression network features a powerful one-pass training procedure and its learning is statistically consistent. The major advantage of utilizing the N-tuple architecture as a regression estimator is the fact that in this realization the training set points are stored by the network implicitly, rather than explicitly, and thus the operation speed remains constant and independent of the training set size. Therefore, the network performance can be guaranteed in practical implementations. Copyright 1996 Elsevier Science Ltd