Improving deep convultional neural networks with unsupervised feature learning

The latest generation of Deep Convolutional Neural Networks (DCNN) have dramatically advanced challenging computer vision tasks, especially in object detection and object classification, achieving state-of-the-art performance in several computer vision tasks including text recognition, sign recognition, face recognition and scene understanding. The depth of these supervised networks has enabled learning deeper and hierarchical representation of features. In parallel, unsupervised deep learning such as Convolutional Deep Belief Network (CDBN) has also achieved state-of-the-art in many computer vision tasks. However, there is very limited research on jointly exploiting the strength of these two approaches. In this paper, we investigate the learning capability of both methods. We compare the output of individual layers and show that many learnt filters and outputs of the corresponding level layer are almost similar for both approaches. Stacking the DCNN on top of unsupervised layers or replacing layers in the DCNN with the corresponding learnt layers in the CDBN can improve the recognition/classification accuracy and training computational expense. We demonstrate the validity of the proposal on ImageNet dataset.

Channel selection in the short-time modulation domain for distant speech recognition

Automatic speech recognition from multiple distant micro- phones poses significant challenges because of noise and reverberations. The quality of speech acquisition may vary between microphones because of movements of speakers and channel distortions. This paper proposes a channel selection approach for selecting reliable channels based on selection criterion operating in the short-term modulation spectrum domain. The proposed approach quantifies the relative strength of speech from each microphone and speech obtained from beamforming modulations. The new technique is compared experimentally in the real reverb conditions in terms of perceptual evaluation of speech quality (PESQ) measures and word error rate (WER). Overall improvement in recognition rate is observed using delay-sum and superdirective beamformers compared to the case when the channel is selected randomly using circular microphone arrays.

Nanostructured hydroxyapatite surfaces-mediated adsorption alters recognition of BMP receptor IA and bioactivity of bone morphogenetic protein-2

Highly efficient loading of bone morphogenetic protein-2 (BMP-2) onto carriers with desirable performance is still a major challenge in the field of bone regeneration. Till now, the nanoscaled surface-induced changes of the structure and bioactivity of BMP-2 remains poorly understood. Here, the effect of nanoscaled surface on the adsorption and bioactivity of BMP-2 was investigated with a series of hydroxyapatite surfaces (HAPs): HAP crystal-coated surface (HAP), HAP crystal-coated polished surface (HAP-Pol), and sintered HAP crystal-coated surface (HAP-Sin). The adsorption dynamics of recombinant human BMP-2 (rhBMP-2) and the accessibility of the binding epitopes of adsorbed rhBMP-2 for BMP receptors (BMPRs) were examined by a quartz crystal microbalance with dissipation. Moreover, the bioactivity of adsorbed rhBMP-2 and the BMP-induced Smad signaling were investigated with C2C12 model cells. A noticeably high mass-uptake of rhBMP-2 and enhanced recognition of BMPR-IA to adsorbed rhBMP-2 were found on the HAP-Pol surface. For the rhBMP-2-adsorbed HAPs, both ALP activity and Smad signaling increased in the order of HAP-Sin < HAP < HAP-Pol. Furthermore, hybrid molecular dynamics and steered molecular dynamics simulations validated that BMP-2 tightly anchored on the HAP-Pol surface with a relative loosened conformation, but the HAP-Sin surface induced a compact conformation of BMP-2. In conclusion, the nanostructured HAPs can modulate the way of adsorption of rhBMP-2, and thus the recognition of BMPR-IA and the bioactivity of rhBMP-2. These findings can provide insightful suggestions for the future design and fabrication of rhBMP-2-based scaffolds/implants.

Specific IgE recognition of pollen allergens from subtropic grasses in patients from the subtropics

Background Pollens of subtropical grasses, Bahia (Paspalum notatum), Johnson (Sorghum halepense), and Bermuda (Cynodon dactylon), are common causes of respiratory allergies in subtropical regions worldwide. Objective To evaluate IgE cross-reactivity of grass pollen (GP) found in subtropical and temperate areas. Methods Case and control serum samples from 83 individuals from the subtropical region of Queensland were tested for IgE reactivity with GP extracts by enzyme-linked immunosorbent assay. A randomly sampled subset of 21 serum samples from patients with subtropical GP allergy were examined by ImmunoCAP and cross-inhibition assays. Results Fifty-four patients with allergic rhinitis and GP allergy had higher IgE reactivity with P notatum and C dactylon than with a mixture of 5 temperate GPs. For 90% of 21 GP allergic serum samples, P notatum, S halepense, or C dactylon specific IgE concentrations were higher than temperate GP specific IgE, and GP specific IgE had higher correlations of subtropical GP (r = 0.771-0.950) than temperate GP (r = 0.317-0.677). In most patients (71%-100%), IgE with P notatum, S halepense, or C dactylon GPs was inhibited better by subtropical GP than temperate GP. When the temperate GP mixture achieved 50% inhibition of IgE with subtropical GP, there was a 39- to 67-fold difference in concentrations giving 50% inhibition and significant differences in maximum inhibition for S halepense and P notatum GP relative to temperate GP. Conclusion Patients living in a subtropical region had species specific IgE recognition of subtropical GP. Most GP allergic patients in Queensland would benefit from allergen specific immunotherapy with a standardized content of subtropical GP allergens.

Exploring accuracy factors in cost estimating practice towards implementing Building Information Modelling (BIM)

Cost estimating has been acknowledged as a crucial component of construction projects. Depending on available information and project requirements, cost estimates evolve in tandem with project lifecycle stages; conceptualisation, design development, execution and facility management. The premium placed on the accuracy of cost estimates is crucial to producing project tenders and eventually in budget management. Notwithstanding the initial slow pace of its adoption, Building Information Modelling (BIM) has successfully addressed a number of challenges previously characteristic of traditional approaches in the AEC, including poor communication, the prevalence of islands of information and frequent reworks. Therefore, it is conceivable that BIM can be leveraged to address specific shortcomings of cost estimation. The impetus for leveraging BIM models for accurate cost estimation is to align budgeted and actual cost. This paper hypothesises that the accuracy of BIM-based estimation, as more efficient, process-mirrors of traditional cost estimation methods, can be enhanced by simulating traditional cost estimation factors variables. Through literature reviews and preliminary expert interviews, this paper explores the factors that could potentially lead to more accurate cost estimates for construction projects. The findings show numerous factors that affect the cost estimates ranging from project information and its characteristic, project team, clients, contractual matters, and other external influences. This paper will make a particular contribution to the early phase of BIM-based project estimation.

Multiview gait biometrics for human identity recognition

We propose a novel multiview fusion scheme for recognizing human identity based on gait biometric data. The gait biometric data is acquired from video surveillance datasets from multiple cameras. Experiments on publicly available CASIA dataset show the potential of proposed scheme based on fusion towards development and implementation of automatic identity recognition systems.

What triggers problem recognition? An exploration on young Australian male problematic online gamers

Help-seeking is a complex decision-making process that first begins with problem recognition. However, little is understood about the conceptualisation of the helpseeking process and the triggers of problem recognition. This research proposes the use of the Critical Incident Technique (CIT) to examine and classify incidents that serve as key triggers of problem recognition among young Australian male problematic online gamers. The research provides a classification of five different types of triggers that will aid social marketers into developing effective early detection, prevention and treatment focused social marketing interventions.

On robust face recognition via sparse coding: The good, the bad and the ugly

In the field of face recognition, sparse representation (SR) has received considerable attention during the past few years, with a focus on holistic descriptors in closed-set identification applications. The underlying assumption in such SR-based methods is that each class in the gallery has sufficient samples and the query lies on the subspace spanned by the gallery of the same class. Unfortunately, such an assumption is easily violated in the face verification scenario, where the task is to determine if two faces (where one or both have not been seen before) belong to the same person. In this study, the authors propose an alternative approach to SR-based face verification, where SR encoding is performed on local image patches rather than the entire face. The obtained sparse signals are pooled via averaging to form multiple region descriptors, which then form an overall face descriptor. Owing to the deliberate loss of spatial relations within each region (caused by averaging), the resulting descriptor is robust to misalignment and various image deformations. Within the proposed framework, they evaluate several SR encoding techniques: l1-minimisation, Sparse Autoencoder Neural Network (SANN) and an implicit probabilistic technique based on Gaussian mixture models. Thorough experiments on AR, FERET, exYaleB, BANCA and ChokePoint datasets show that the local SR approach obtains considerably better and more robust performance than several previous state-of-the-art holistic SR methods, on both the traditional closed-set identification task and the more applicable face verification task. The experiments also show that l1-minimisation-based encoding has a considerably higher computational cost when compared with SANN-based and probabilistic encoding, but leads to higher recognition rates.

Conceptualising recognition of prior learning processes in the age of open learning

This chapter interrogates what recognition of prior learning (RPL) can and does mean in the higher education sector—a sector in the grip of the widening participation agenda and an open access age. The chapter discusses how open learning is making inroads into recognition processes and examines two studies in open learning recognition. A case study relating to e-portfolio-style RPL for entry into a Graduate Certificate in Policy and Governance at a metropolitan university in Queensland is described. In the first instance, candidates who do not possess a relevant Bachelor degree need to demonstrate skills in governmental policy work in order to be eligible to gain entry to a Graduate Certificate (at Australian Qualifications Framework Level 8) (Australian Qualifications Framework Council, 2013, p. 53). The chapter acknowledges the benefits and limitations of recognition in open learning and those of more traditional RPL, anticipating future developments in both (or their convergence).

Acoustic classification of Australian anurans using syllable features

Acoustic classification of anurans (frogs) has received increasing attention for its promising application in biological and environment studies. In this study, a novel feature extraction method for frog call classification is presented based on the analysis of spectrograms. The frog calls are first automatically segmented into syllables. Then, spectral peak tracks are extracted to separate desired signal (frog calls) from background noise. The spectral peak tracks are used to extract various syllable features, including: syllable duration, dominant frequency, oscillation rate, frequency modulation, and energy modulation. Finally, a k-nearest neighbor classifier is used for classifying frog calls based on the results of principal component analysis. The experiment results show that syllable features can achieve an average classification accuracy of 90.5% which outperforms Mel-frequency cepstral coefficients features (79.0%).

Acoustic feature extraction using perceptual wavelet packet decomposition for frog call classification

Frog protection has become increasingly essential due to the rapid decline of its biodiversity. Therefore, it is valuable to develop new methods for studying this biodiversity. In this paper, a novel feature extraction method is proposed based on perceptual wavelet packet decomposition for classifying frog calls in noisy environments. Pre-processing and syllable segmentation are first applied to the frog call. Then, a spectral peak track is extracted from each syllable if possible. Track duration, dominant frequency and oscillation rate are directly extracted from the track. With k-means clustering algorithm, the calculated dominant frequency of all frog species is clustered into k parts, which produce a frequency scale for wavelet packet decomposition. Based on the adaptive frequency scale, wavelet packet decomposition is applied to the frog calls. Using the wavelet packet decomposition coefficients, a new feature set named perceptual wavelet packet decomposition sub-band cepstral coefficients is extracted. Finally, a k-nearest neighbour (k-NN) classifier is used for the classification. The experiment results show that the proposed features can achieve an average classification accuracy of 97.45% which outperforms syllable features (86.87%) and Mel-frequency cepstral coefficients (MFCCs) feature (90.80%).

Application of image processing techniques for frog call classification

Frogs have received increasing attention due to their effectiveness for indicating the environment change. Therefore, it is important to monitor and assess frogs. With the development of sensor techniques, large volumes of audio data (including frog calls) have been collected and need to be analysed. After transforming the audio data into its spectrogram representation using short-time Fourier transform, the visual inspection of this representation motivates us to use image processing techniques for analysing audio data. Applying acoustic event detection (AED) method to spectrograms, acoustic events are firstly detected from which ridges are extracted. Three feature sets, Mel-frequency cepstral coefficients (MFCCs), AED feature set and ridge feature set, are then used for frog call classification with a support vector machine classifier. Fifteen frog species widely spread in Queensland, Australia, are selected to evaluate the proposed method. The experimental results show that ridge feature set can achieve an average classification accuracy of 74.73% which outperforms the MFCCs (38.99%) and AED feature set (67.78%).

Acoustic features for multi-level classification of Australian frogs

Over past few decades, frog species have been experiencing dramatic decline around the world. The reason for this decline includes habitat loss, invasive species, climate change and so on. To better know the status of frog species, classifying frogs has become increasingly important. In this study, acoustic features are investigated for multi-level classification of Australian frogs: family, genus and species, including three families, eleven genera and eighty five species which are collected from Queensland, Australia. For each frog species, six instances are selected from which ten acoustic features are calculated. Then, the multicollinearity between ten features are studied for selecting non-correlated features for subsequent analysis. A decision tree (DT) classifier is used to visually and explicitly determine which acoustic features are relatively important for classifying family, which for genus, and which for species. Finally, a weighted support vector machines (SVMs) classifier is used for the multi- level classification with three most important acoustic features respectively. Our experiment results indicate that using different acoustic feature sets can successfully classify frogs at different levels and the average classification accuracy can be up to 85.6%, 86.1% and 56.2% for family, genus and species respectively.

Sampling accuracy of reef resource inventory technique

A new technique called the reef resource inventory (RRI) was developed to map the distribution and abundance of benthos and substratum on reefs. The rapid field sampling technique uses divers to visually estimate the percentage cover of categories of benthos and substratum along 2x20 in plotless strip-transects positioned randomly over the tops, and systematically along the edge of reefs. The purpose of this study was to compare the relative sampling accuracy of the RRI against the line intercept transect technique (LIT), an international standard for sampling reef benthos and substratum. Analysis of paired sampling with LIT and RRI at 51 sites indicated sampling accuracy was not different (P > 0.05) for 8 of the 12 benthos and substratum categories used in the study. Significant differences were attributed to small-scale patchiness and cryptic coloration of some benthos; effects associated with sampling a sparsely distributed animal along a line versus an area; difficulties in discriminating some of the benthos and substratum categories; and differences due to visual acuity since LIT measurements were taken by divers close to the seabed whereas RRI measurements were taken by divers higher in the water column. The relative cost efficiency of the RRI technique was at least three times that of LIT for all benthos and substratum categories and as much as 10 times higher for two categories. These results suggest that the RRI can be used to obtain reliable and accurate estimates of relative abundance of broad categories of reef benthos and substratum.

Repeatable Condition-Invariant Visual Odometry for Sequence-Based Place Recognition

This paper describes a vision-only system for place recognition in environments that are tra- versed at different times of day, when chang- ing conditions drastically affect visual appear- ance, and at different speeds, where places aren’t visited at a consistent linear rate. The ma- jor contribution is the removal of wheel-based odometry from the previously presented algo- rithm (SMART), allowing the technique to op- erate on any camera-based device; in our case a mobile phone. While we show that the di- rect application of visual odometry to our night- time datasets does not achieve a level of perfor- mance typically needed, the VO requirements of SMART are orthogonal to typical usage: firstly only the magnitude of the velocity is required, and secondly the calculated velocity signal only needs to be repeatable in any one part of the environment over day and night cycles, but not necessarily globally consistent. Our results show that the smoothing effect of motion constraints is highly beneficial for achieving a locally consis- tent, lighting-independent velocity estimate. We also show that the advantage of our patch-based technique used previously for frame recogni- tion, surprisingly, does not transfer to VO, where SIFT demonstrates equally good performance. Nevertheless, we present the SMART system us- ing only vision, which performs sequence-base place recognition in extreme low-light condi- tions where standard 6-DOF VO fails and that improves place recognition performance over odometry-less benchmarks, approaching that of wheel odometry.