Low-cost hardware speech enhancement for improved speech recognition in automotive environments

Voice recognition is one of the key enablers to reduce driver distraction as in-vehicle systems become more and more complex. With the integration of voice recognition in vehicles, safety and usability are improved as the driver’s eyes and hands are not required to operate system controls. Whilst speaker independent voice recognition is well developed, performance in high noise environments (e.g. vehicles) is still limited. La Trobe University and Queensland University of Technology have developed a low-cost hardware-based speech enhancement system for automotive environments based on spectral subtraction and delay–sum beamforming techniques. The enhancement algorithms have been optimised using authentic Australian English collected under typical driving conditions. Performance tests conducted using speech data collected under variety of vehicle noise conditions demonstrate a word recognition rate improvement in the order of 10% or more under the noisiest conditions. Currently developed to a proof of concept stage there is potential for even greater performance improvement.

Palliative care curriculum for speech–language pathology students

This paper reports on the experience of undergraduate speech–language pathology students at one university chosen for the implementation stage of the Palliative Care Curriculum for Undergraduates (PCC4U) Project. Funded by a government department for health and ageing through a national palliative care programme, the project was managed by a team of researchers from the discipline of nursing. The PCC4U project championed the inclusion of palliative care education as an integral part of medical, nursing, and allied healthcare undergraduate training. Of the pilot sites chosen for the PCC4U project, only one site, reported here, included both speech–language pathology and social work disciplines, providing an important opportunity for interdisciplinary collaboration on novel curriculum development in an area of mutual interest. This synergy served as an excellent foundation for ongoing opportunities for interdisciplinary teaching and learning in the university. Speech–language pathology students reported that the project was an invaluable addition to their education and preparation for clinical practice.

Grade A Taipei junior high school principals’ self-reported leadership practices

The principal’s leadership and curriculum development are considered the core elements for creating a high performing junior high school. In Taiwan, mathematics curriculum reform has been an ongoing topic since 1994. The pedagogy, classroom interactions, and the underlying philosophy of mathematics education have varied with different versions of guidelines. These changes inevitably increase the requirement for principals’ leadership in order to effectively implement the curriculum reform. Principals’ leadership is essential to the success of the implementation in their school. This study aimed to explore and identify the leadership of junior high school principals whose schools had been judged by the Taipei City Government as Grade A junior high schools. Principals’ implementations of the reformed mathematics curriculum were used as examples to generate insights of their leadership. This study drew upon a multiple-case study approach. Data were collected from interviews, observations, and documentations. Bass and Avolio’s (1997) full range leadership theory provided a structure for gaining insight into these principals’ leadership practices. Five Grade A Taipei junior high school principals participated and shared their leadership concepts and experiences. Findings revealed that the leadership preferences of the five principles varied considerably. Management by exception-active, contingent reward, individualised consideration, and idealised influence were Grade A Taipei junior high school principals’ preferred leadership practices. In addition, principals’ leadership strategies associated with these practices were identified. These principals had adopted a range of leadership strategies according to the staff and school needs. Results of this study have implications for both Taiwanese principals and education departments. Principals can enhance their leadership by gaining more understanding about the Grade A principals’ leadership practices and strategies. Taiwanese education departments can improve school leadership training programs by focusing on these practices and strategies, which may also lead to more effective strategies for implementing national curriculum reform.

The prevalence and characteristics of self-reported dangerous driving behaviours among a young cohort

Background: Young motorists engaging in anti-social and often dangerous driving manoeuvres (which is often referred to as “hooning” within Australia) is an increasing road safety problem. While anecdotal evidence suggests that such behaviour is positively linked with crash involvement, researchers have yet to examine whether younger drivers who deliberately break road rules and drive in an erratic manner (usually with peers) are in fact over represented in crash statistics. This paper outlines research that aimed to identify the characteristics of individuals most likely to engaging in hooning behaviours, as well as examine the frequency of such driving behaviours and if such activity is linked with self-reported crash involvement.---------- Methods: A total of 717 young drivers in Queensland voluntarily completed a questionnaire to investigate their driving behaviour and crash history.---------- Results: Quantitative analysis of the data revealed that almost half the sample reported engaging in some form of “hooning” behaviour at least once in their lifetime, although only 4% indicated heavy participation in the behaviour e.g., >50 times. Street racing was the most common activity reported by participants followed by “drifting” and then “burnouts”. Logistic regression analysis indicated that being younger and a male was predictive of reporting such anti-social driving behaviours, and importantly, a trend was identified between such behaviour and self-reported crash involvement.---------- Conclusions: This research provides preliminary evidence that younger male drivers are more likely to engage in dangerous driving behaviours, which ultimately may prove to increase their overall risk of becoming involved in a crash. This paper will further outline the study findings in regards to current enforcement efforts to deter such driving activity as well as provide direction for future research efforts in this area.---------- Research highlights: ► The self-reported driving behaviours of 717 younger Queensland drivers were examined to investigate the relationship between deliberately breaking road rules and self-reported crash involvement. ► Younger male drivers were most likely to engage in such aberrant driving behaviours and a trend was identified between such behaviour and self-reported crash involvement.

Recognising audio-visual speech in vehicles using the AVICAR database

Interacting with technology within a vehicle environment using a voice interface can greatly reduce the effects of driver distraction. Most current approaches to this problem only utilise the audio signal, making them susceptible to acoustic noise. An obvious approach to circumvent this is to use the visual modality in addition. However, capturing, storing and distributing audio-visual data in a vehicle environment is very costly and difficult. One current dataset available for such research is the AVICAR [1] database. Unfortunately this database is largely unusable due to timing mismatch between the two streams and in addition, no protocol is available. We have overcome this problem by re-synchronising the streams on the phone-number portion of the dataset and established a protocol for further research. This paper presents the first audio-visual results on this dataset for speaker-independent speech recognition. We hope this will serve as a catalyst for future research in this area.

Bayes Factor based speaker segmentation for speaker diarization

This paper proposes the use of the Bayes Factor as a distance metric for speaker segmentation within a speaker diarization system. The proposed approach uses a pair of constant sized, sliding windows to compute the value of the Bayes Factor between the adjacent windows over the entire audio. Results obtained on the 2002 Rich Transcription Evaluation dataset show an improved segmentation performance compared to previous approaches reported in literature using the Generalized Likelihood Ratio. When applied in a speaker diarization system, this approach results in a 5.1% relative improvement in the overall Diarization Error Rate compared to the baseline.

Self-reported driving difficulty by persons with Hemianopia and Quadrantanopia

Purpose: To compare self-reported driving difficulty by persons with hemianopic or quadrantanopic field loss with that reported by age-matched drivers with normal visual fields; and to examine how their self- reported driving difficulty compares to ratings of driving performance provided by a certified driving rehabilitation specialist(CDRS). Method: Participants were 17 persons with hemianopic field loss, 7 with quadrantanopic loss, and 24 age-matched controls with normal visual fields, all of whom had current drivers’ licenses. Information was collected via questionnaire regarding driving difficulties experienced in 21 typical driving situations grouped into 3 categories(involvement of peripheral vision, low visibility conditions, and independent mobility). On-road driving performance was evaluated by a CDRS using a standard assessment scale. Results: Drivers with hemianopic and quadrantanopic field loss expressed significantly more difficulty with driving maneuvers involving peripheral vision and independent mobility, compared to those with normal visual fields. Drivers with hemianopia and quadrantanopia who were rated as unsafe to drive based upon an on-road assessment by the CDRS were no more likely to report driving difficulty than those rated as safe. Conclusion: This study highlights aspects of driving that hemianopic or quadrantanopic persons find particularly problematic, thus suggesting areas that could be focused on driving rehabilitation. Some drivers with hemianopia or quadrantanopia may inappropriately view themselves as good drivers when in fact their driving performance is unsafe as judged by a driving professional.

Improving visual noise insensitivity in small vocabulary audio visual speech recognition applications

Visual noise insensitivity is important to audio visual speech recognition (AVSR). Visual noise can take on a number of forms such as varying frame rate, occlusion, lighting or speaker variabilities. The use of a high dimensional secondary classifier on the word likelihood scores from both the audio and video modalities is investigated for the purposes of adaptive fusion. Preliminary results are presented demonstrating performance above the catastrophic fusion boundary for our confidence measure irrespective of the type of visual noise presented to it. Our experiments were restricted to small vocabulary applications.

Speech recognition in adverse environments using lip information

The performance of automatic speech recognition systems deteriorates in the presence of noise. One known solution is to incorporate video information with an existing acoustic speech recognition system. We investigate the performance of the individual acoustic and visual sub-systems and then examine different ways in which the integration of the two systems may be performed. The system is to be implemented in real time on a Texas Instruments' TMS320C80 DSP.

Robust speaker verification via fusion of speech and lip modalities

This paper investigates the use of lip information, in conjunction with speech information, for robust speaker verification in the presence of background noise. It has been previously shown in our own work, and in the work of others, that features extracted from a speaker's moving lips hold speaker dependencies which are complementary with speech features. We demonstrate that the fusion of lip and speech information allows for a highly robust speaker verification system which outperforms the performance of either sub-system. We present a new technique for determining the weighting to be applied to each modality so as to optimize the performance of the fused system. Given a correct weighting, lip information is shown to be highly effective for reducing the false acceptance and false rejection error rates in the presence of background noise

The use of temporal speech and lip information for multi-modal speaker identification via multi-stream HMMs

Investigates the use of temporal lip information, in conjunction with speech information, for robust, text-dependent speaker identification. We propose that significant speaker-dependent information can be obtained from moving lips, enabling speaker recognition systems to be highly robust in the presence of noise. The fusion structure for the audio and visual information is based around the use of multi-stream hidden Markov models (MSHMM), with audio and visual features forming two independent data streams. Recent work with multi-modal MSHMMs has been performed successfully for the task of speech recognition. The use of temporal lip information for speaker identification has been performed previously (T.J. Wark et al., 1998), however this has been restricted to output fusion via single-stream HMMs. We present an extension to this previous work, and show that a MSHMM is a valid structure for multi-modal speaker identification

The use of speech and lip modalities for robust speaker verification under adverse conditions

Investigates the use of lip information, in conjunction with speech information, for robust speaker verification in the presence of background noise. We have previously shown (Int. Conf. on Acoustics, Speech and Signal Proc., vol. 6, pp. 3693-3696, May 1998) that features extracted from a speaker's moving lips hold speaker dependencies which are complementary with speech features. We demonstrate that the fusion of lip and speech information allows for a highly robust speaker verification system which outperforms either subsystem individually. We present a new technique for determining the weighting to be applied to each modality so as to optimize the performance of the fused system. Given a correct weighting, lip information is shown to be highly effective for reducing the false acceptance and false rejection error rates in the presence of background noise